Changes in the Building Stock of Da Nang Between 2015 and 2017

data

Data Descriptor Changes in the Building Stock of Da Nang between 2015 and 2017

Andreas Braun 1,* , Gebhard Warth 1, Felix Bachofer 2 , Tram Thi Quynh Bui 3, Hao Tran 4 and Volker Hochschild 1 1 Institute of Geography, University of Tübingen, 72070 Tübingen, Germany; [email protected] (G.W.); [email protected] (V.H.) 2 German Aerospace Center (DLR), Earth Observation Center (EOC), 82334 Wessling, Germany; [email protected] 3 Da Nang Institute for Socio-Economic Development (DISED), Da Nang City 550000, Vietnam; [email protected] 4 Da Nang Urban Planning Institute (UPI), Da Nang City 550000, Vietnam; [email protected] * Correspondence: [email protected]; Tel.: +49-7071-297-8940

Received: 16 March 2020; Accepted: 20 April 2020; Published: 23 April 2020

Abstract: This descriptor introduces a novel dataset, which contains the number and types of buildings in the city of Da Nang in Central Vietnam. The buildings were classified into nine distinct types and initially extracted from a satellite image of the year 2015. Secondly, changes were identified based on a visual interpretation of an image of the year 2017, so that new buildings, demolished buildings and building upgrades can be quantitatively analyzed. The data was aggregated by administrative wards and a hexagonal grid with a diameter of 250 m to protect personal rights and to avoid the misuse of a single building’s information. The dataset shows an increase of 19,391 buildings between October 2015 and August 2017, with a variety of interesting spatial patterns. The center of the city is mostly dominated by building changes and upgrades, while most of the new buildings were constructed within a distance of five to six kilometers from the city center.

Dataset: https://doi.org/10.5281/zenodo.3757710

Dataset License: CC-BY 4.0

Keywords: building types; change detection; remote sensing; urbanization; urban growth; Southeast Asia

1. Summary Rapidly urbanizing cities require reliable and up-to-date information about their inhabitants to ensure the provision of food and energy, but also for the planning of logistics, public services and infrastructure. However, the collection of this information by surveys or an official census is time-consuming and expensive. As an alternative, methods of earth observation can be employed to extract the number of buildings as an indicator of the size and spatial distribution of the urban population [1,2]. The presented dataset provides building types and numbers for the city of Da Nang for the years 2015 and 2017. The initial number of building types was retrieved from a very high-resolution (VHR) Pléiades satellite image using an object-based image analysis (OBIA) approach, as was already described in two previous studies by the authors [3,4]. Results from these studies were then refined by the approach presented in this data descriptor to extract the changes between 2015 and 2017 based on visual inspection and the identification of changes, classified into five different categories: no change, new building, changed building, upgraded building and demolished building.

Data 2020, 5, 42; doi:10.3390/data5020042 www.mdpi.com/journal/data Data 2020, 5, 42 2 of 18

Data 2020, 5, x FOR PEER REVIEW 2 of 18 The building types and changes were aggregated in two datasets: administrative boundaries and hexagonsThe building of 250 types m in and width. changes These were two aggregated datasets are in the two first datasets: publicly administrative available and boundaries most up to dateand assessmentshexagons of of250 the m buildingin width. stock These of two Da datasets Nang. Such are the information first publicly is valuable available for and urban most planners up to date to estimateassessments the futureof the growthbuilding of stock the city, of theDa Nang demand. Such ofthe information population is (food,valuable energy, for urban or water) planners and the to personsestimate potentially the future agrowthffected of by the rising city, sea the levels demand under of the climate population change. (food, energy, or water) and the persons potentially affected by rising sea levels under climate change. 2. Methods 2. MethodsSeveral steps were required to identify the changes in the urban area of Da Nang, including the collectionSeveral of referencesteps were data, required the retrieval to identify of built-upthe changes areas in andthe urban building area types of Da from Nang satellite, including images the andcollection the visual of reference identification data, the of retrieval changes of over built the-up investigated areas and building period. types These from steps satellite are summarized images and the in Figurevisual1 identification. of changes over the investigated period. These steps are summarized in Figure 1.

FigureFigure 1.1. SystematicSystematic workﬂowworkflow forfor thethe creation of the presented dataset.dataset.

2.1.2.1. StudyStudy AreaArea TheThe studystudy areaarea coverscovers thethe urbanurban andand suburbansuburban partsparts of thethe citycity ofof DaDa Nang,Nang, locatedlocated in Central Vietnam.Vietnam. ItIt hashas aa populationpopulation ofof 1.1251.125 millionmillion inhabitants,inhabitants, with an annualannual growth rate around 2.75% accordingaccording to to the the estimates estimates ofof thethe UnitedUnited NationsNations [[5]5].. Because of its harbor, which is the fourth largest in thethe country, country, the the entirety entirety of of Central Central VietnamVietnam hashas becomebecome economicallyeconomically importantimportant and has experienced aa considerableconsiderable population population increase increase since since 2007, 2007, mostly mostly due due to to migration migration from from rural rural areas areas [[6]6].. The administrativeadministrativearea areacovered covered byby thethe datasetdataset isis divideddivided intointo sevenseven districts and 5050 wards ( (FigureFigure 22,, blackblack lines),lines), ofof which six six districts districts belong belong to tothe the city city of Da of Nang Da Nang.. Because Because the acquisition the acquisition of VHR of data VHR is limited data is limitedby financial by ﬁnancial and technical and technical constraints, constraints, an area of an interest area of (AOI) interest of 225 (AOI) square of 225kilometers square kilometerswas defined was for deﬁnedthe analysis for the (Figure analysis 2, red (Figure line),2 which, red line), covers which the urban covers and the suburban urban and settlement suburban areas settlement of the areascity and of thelarge city parts and of large its rural parts surroundings. of its rural surroundings. Because of the Because seamless of transition the seamless of the transition urban area of at the the urban southern area atborder the southern of the study border area, of parts the study of the area, Điện parts Bàn district of the Đ ofiệ nthe B àadjacentn district Quảng of the Nam adjacent province Quả ngare Nam also provincecovered by are the also AOI covered and thus by thehave AOI been and included thus have in the been presented included dataset. in the presented dataset.

Data 2020, 5, 42 3 of 18 Data 2020, 5, x FOR PEER REVIEW 3 of 18

Figure 2. Study area, extent of the analysis, and field reference data collection. Figure 2. Study area, extent of the analysis, and field reference data collection. 2.2. Satellite Data 2.2. Satellite Data Two image products from the Pléiades satellite constellation acquired on 24.10.2015 and 13.08.2017 were usedTwo image in this products study. For from this the period, Pléiades the satellite United Nationsconstellation estimated acquired a growth on 24.10.2015 of 3.5% and for the13.08.2017 city of Dawere Nang used [ 5in]. this The study sensor. For acquires this period, four the multispectral United Nations bands estimated (blue, green, a growth red, of and 3.5% infrared) for the city at 2.8 of Da m andNan oneg [5] panchromatic. The sensor acquires band atfour 0.7 multispectral m spatial resolution bands (blue, (resampled green, red, to and 50 cm infrared) by the at data 2.8 provider).m and one Bothpanchromatic products band were at pan-sharpened 0.7 m spatial resolution to a final (resampled spatial resolution to 50 cm by of 50the cmdata and provider). radiometrically Both products and geometricallywere pan-sharpe calibratedned to a final as described spatial resolution by Warth of et 50 al. cm [4 ].and The radiometrically two images were and geometrically acquired from calibrated slightly diasff describederent off-nadir by Warth angles, et al. which [4]. The resulted two images in the geometricwere acquired shift from of high slightly buildings different towards off-nadir the sensor.angles, Furthermore,which resulted the in dithefferent geometric acquisition shift of dates high withinbuildings the towards year lead the to sensor. differences Furthe inrmore, illumination the different and theacquisition shadows dates that within are cast the from year large lead objects. to differences Lastly, therein illumination are phenologic and the di ffshadowserences betweenthat are cast October from andlarge August, objects. soLastly, urban there greenspaces are phenologic or trees differences can produce between unequal October patterns and August, in both so images.urban greenspaces All these slightor trees distortions can produce can uneq makeual the patterns visual in identification both images. ofAll changes these slight challenging, distortions especially can make in the densely visual built-upidentification areas. of changes challenging, especially in densely built-up areas. FigureFigure3 3gives gives an an example example on on the the di differencesfferences betweenbetween bothboth imagesimages forfor aa smallsmall selectedselected areaarea inin thethe AnAn HHảiải BBắcắc wardward ofof thethe SSơnơn TrTràà district district in in the the eastern eastern partpart ofof DaDa Nang. Nang. AlthoughAlthough thethe buildingbuilding inin thethe centercenter hashas notnot changedchanged betweenbetween bothboth imageimage acquisitions,acquisitions, thethe redred colorcolor ofof thethe roofroof lookslooks didifferentfferent becausebecause ofof sunlightsunlight illumination.illumination. Moreover,Moreover, thethe shadowsshadows atat thethe northernnorthern edgesedges ofof talltall buildingsbuildings areare onlyonly presentpresent inin thethe imageimage fromfrom OctoberOctober 2015.2015. TheseThese seasonalseasonal didifferencesfferences werewere reducedreduced duringduring thethe radiometricradiometric calibration, calibration but, but could could not entirely not entirely be removed. be removed. Accordingly, Accordingly, these radiometric these radiometric differences candifferences potentially can increasepotentially the increase chance ofthe falsely chance assigned of falsely changes assigned at changes the building at the level building during level the during latter visualthe lat interpretationter visual interpretation of changes atof the changes building at level.the building The figure level. also The demonstrates figure also that demonstrates actual changes that inactual the morphologicchanges in the structure morphologic of the structure city, such of as the the city creation, such as of the paved creation roads o (inf paved the northern roads (in and the westernnorthern part and of western the image), part orof the image), construction or the of construction new buildings of new (two buildings bright lines (two in bright the center lines north) in the orcenter other north) infrastructure or other infrastructure (blue sports court (blue in sports the center) court in can the clearly center) be can identified clearly bybe identified visual comparison. by visual comparison. However, slight differences had to be kept in mind by the interpreters in order to avoid errors of commission during the identification of changes between 2015 and 2017.

Data 2020, 5, 42 4 of 18

However, slight diﬀerences had to be kept in mind by the interpreters in order to avoid errors of commissionData 2020, 5, x FOR during PEER the REVIEW identiﬁcation of changes between 2015 and 2017. 4 of 18

Figure 3.3. Radiometric differences diﬀerences resulting resulting from from angle angle and and acquisition acquisition date date / timetime ofof thethe twotwo imagesimages used in thisthis study.study.

AsAs a qualitativequalitative historic reference, an image acquired on 15.08.1977 by the Keyhole-9Keyhole-9 (KH-9)(KH-9) mission “Hexagon” [[7]7] was used inin thisthis study.study. ItsIts spatialspatial resolutionresolution ofof 77 mm allowed us toto identifyidentify thethe historic citycity centercenter (see(see SectionSection 3.23.2)) ofof DaDa Nang.Nang.

2.3. Reference Data 2.3. Reference Data To gain a better understanding of the city’s morphology and the predominant building types, three To gain a better understanding of the city’s morphology and the predominant building types, field campaigns have been carried out in 2015 and 2016. Using mobile handheld tablet devices, a total three field campaigns have been carried out in 2015 and 2016. Using mobile handheld tablet devices, number of 803 geocoded questionnaires on single buildings were collected, including information on a total number of 803 geocoded questionnaires on single buildings were collected, including building size, shape and height, its usage, the color and materials of the roof, photographs and further information on building size, shape and height, its usage, the color and materials of the roof, parameters regarding the structure of the buildings. These datasets were collected in March 2015 photographs and further parameters regarding the structure of the buildings. These datasets were n n n (collected= 233), in March March 2016 2015 ( (=n 465)= 233), and March December 2016 2016(n = 465) ( = and105). December Based on 2016 the collected (n = 105). data, Based building on the archetypescollected data, were building defined archetypes in close collaboration were defined with in close local collaboration experts. The with spatial local distribution experts. The of spatial these buildingsdistribution and of the these assigned buildings building and the type assigned are shown building in Figure type1. Typically,are shown in in Vietnamese Figure 1. Typical cities, shoply, in housesVietnamese with acities, longish shop shape houses and with three a tolongish five floors shape of and mixed three use to are five the floors most of common mixed use in the are city. the Theymost cancommon be refined in the into city. sub-classes They can (such be refined as traditional, into sub new,-clas commercial,ses (such as rowhouse), traditional, as initially new, commercial, suggested byrowhouse), Moon et al.as initially (2009) [8 suggested], but only by with Moon additional et al. (2009) knowledge [8], but about only with the degree additional of commercial knowledge use about and constructionthe degree of materials, commercial which use cannot and construction be retrieved materials, from remote which sensing cannot data be for retrieved the entire from city. remote 2.4.sensing Building data Typologyfor the entire city.

2.4. BuildingThe building Typology typology was developed based on an existing study by Downes et al. (2016) for Ho Chi Minh City [9]. It has already been applied to estimate the number of diﬀerent fractions of The building typology was developed based on an existing study by Downes et al. (2016) for Ho household waste [3] and was successfully transferred to the Rwandan capital, Kigali, to demonstrate Chi Minh City [9]. It has already been applied to estimate the number of different fractions of the spatial patterns of urbanization between 2008 and 2015 [10]. The criteria for the building types are household waste [3] and was successfully transferred to the Rwandan capital, Kigali, to demonstrate based on height, size, shape and roof material to allow an assignment of building types for the entire the spatial patterns of urbanization between 2008 and 2015 [10]. The criteria for the building types urban area of Da Nang, solely based on satellite image information. Table1 lists the nine building are based on height, size, shape and roof material to allow an assignment of building types for the types, together with a short description of their identiﬁcation from remote sensing images. entire urban area of Da Nang, solely based on satellite image information. Table 1 lists the nine building types, together with a short description of their identification from remote sensing images.

DataData2020 2020, 5,, 5 42, x FOR PEER REVIEW 5 of5 of1818 Data 2020, 5, x FOR PEER REVIEW 5 of 18 Data 2020, 5, x FOR PEER REVIEW 5 of 18 Table 1. Building archetypes for the City of Da Nang, adapted from Vetter Gindele et al. (2019) [3.] Table 1. Building archetypes for the City of Da Nang, adapted from Vetter Gindele et al. (2019) [3.] TableTable 1. 1.Building Building archetypes archetypes forfor thethe CityCity ofof DaDa Nang,Nang, adapted from from Vetter Vetter Gindele Gindele et et al. al. (2019) (2019) [3 [.3] ]. Building Type and ID Building Type and Description Representative Reference Picture IDID BuildingBuildingStatistics Type Type and and Statistics Description Description RepresentativeRepresentative Reference Picture Picture ID Statistics Description Representative Reference Picture Single family,Statistics basic (n = 65) Single family, basic (n = 65) Single family, basic (n = 65) Detached housing, with a mix of Single family, basic (n = 65)Detached housing, with a mix of residential average (standard deviation) Detachedresidential housing, with and commerciala mix of residential use. averageaverage (standard (standard deviation) deviation) Detached of:and commercial housing, use. with Low a mix-rise of with residential one to average (standardof: deviation) and commercialLow-rise withuse. Low one-rise to two with floors. one to height:of: 2.3 m (2.6 m) twoand floors. commercial Comprised use. Low of wood,-rise with brickwork, one to 1 height: 2.3of: m (2.6 m) two floors. Comprised of wood, brickwork, 1 height:1 2.3 m (2.62 m) 2 twoand floors. Comprisedreinforced Comprised concrete, of wood,of wood, tin brickwork,roof. brickwork, Often 11 height:size1 : 57.72.3 m m (2.6(39.1 m) m ) and reinforced concrete, tin roof. Often size1: 57.7 m² (39.1 m²) and reinforced concrete, tin roof. Often size1: 57.71 m² (39.1 m²) and andreinforced reinforced concrete, concrete, tin roof. tin Often roof. sizelength1: 57.71 : m² 16.5 (39.1 m (13.2m²) m) located along small alleyways or in peri- lengthsize : 57.7: 16.5 m² m (39.1 (13.2 m²) m) located along small alleyways or in peri- 1 1 located alongOften small located alleyways along or small in peri- length1: 16.5 m (13.2 m) urban locations. lengthwidthwidth:1 :16.5 8.0: 8.0mm (7.2(13.2 m (7.2m) m) m) urban locations. 1 alleywaysurban locations. or in peri-urban width1: 8.0 m (7.2 m)2 persons/householdpersonswidth/household: 8.0 m (7.22: 4.3 m): (0.9) 4.3 (0.9) 2 locations. persons/household2: 4.3 (0.9) persons/householdLocal - type shophouse: 4.3 (0.9) Local - type shophouse Local - type(n = 540) shophouse (n = 540) TypicalTypical building building type in type Da Nang. in Da Local-type(n = 540) shophouse Typical building type in Da Nang. Detached/semiTypicalNang. building- Detached detached/terraced type in Da/semi- Nang. average (standard(n = 540) deviation) Detached/semidetached/terraced average (standard deviation) shophouse.Detached/semidetached It is a/ terracedtwo- detached/terraced to five shophouse. story urban averageaverage (standard (standardof: deviation) deviation) of:shophouse. It is a two to five story urban 2 of: building,shophouse.It which is aIt twois allowsa two to five tofor five storya shop story urban or urban other 2 height:height: 3.9of: 3.9m (3.1 m (3.1m) m) building, which allows for a shop or other 22 height:1 3.9 m (3.12 m) 2 building,publicbuilding, activity which allows at whichthe streetforallows a shoplevel, or for with other a sizeheight:size1: 85.1: 85.13.9 m² m m(60.3 (3.1(60.3 m)m²) m ) public activity at the street level, with 1 public activity at the street level, with size1: 85.11 m² (60.3 m²) residentialshop accommodation or other public on activity the upper at lengthsizelength: 85.11: 16.4 :m² 16.4 m (60.3 (7.0 m m²) (7.0m) m) residential accommodation on the upper 1 residential accommodation on the upper length1: 16.41 m (7.0 m) the street level,floors. with residential lengthwidthwidth1:: 16.47.2: 7.2m m (4.2 (7.0 m (4.2m) m) m) floors. 1 floors. width1: 7.2 m (4.2 m) accommodation on the upper persons/householdpersonswidth/household: 7.2 m (4.22: 5.8 m)2: (2.4) 5.8 (2.4) 2 persons/household2: 5.8 (2.4) floors. persons/householdSingle family, bungalow: 5.8 (2.4) Single family, bungalow Single family,(n = 29) bu ngalow (n = 29) Single( family,n = 29) bungalow (n = 29) average (standard deviation) Single family detached dwelling, averageaverage (standard (standard deviation) deviation) of:Single family detached dwelling, low-rise, average (standardof: deviation) Single familylow-rise, detached two todwelling, three floors, low-rise, 3 height:of: 4.0 m (1.5 m) Singletwo to family three floors,detached built dwelling, from bric lowkwork-rise , 33 height: 4.0of: m (1.5 m) two to threebuilt floors, from built brickwork from bric andkwork 3 height:1 4.0 m (1.52 m) 2 andtwo concrete, to three located floors, builtin new from urban bric districts.kwork sizeheight:1 : 101.6 4.0 m m(1.5(55.4 m) m ) and concrete, located in new urban districts. size1: 101.6 m² (55.4 m²) and concrete,concrete, located located in new in urban new urbandistricts. size1: 101.61 m² (55.4 m²) and concrete, located in new urban districts. sizelength1: 101.61 : 25.1m² (55.4 m (16.8 m²) m) length1: 25.1 m (16.8 m) districts. length1: 25.11 m (16.8 m) lengthwidth11: 25.1: 14.5 m (16.8 m (9.1 m) m) width1: 14.5 m (9.1 m) width1: 14.5 m (9.1 2m) personswidth1/:household 14.5 m (9.12 m): 4.7 (2.4) persons/household2: 4.7 (2.4) persons/household2: 4.7 (2.4) persons/household2: 4.7 (2.4) Single family, villa Single family, villa Single( nfamily, = 39) villa (n = 39) Single(n = family,39) villa MostlyMostly a single a-family single-family detached detached dwelling, Mostly a single-family detached dwelling, average (standard(n = 39)deviation) Mostlysometimes adwelling, single- familytwo sometimes to detachedthree attached twodwelling, to average (standard deviation) sometimes two to three attached averageaverage (standard (standardof: deviation) deviation) of: sometimesmultifamilythree attached two buildings, to three multifamily lowattached rise, 4 height:of: 5.1 m (3.7 m) multifamilybuildings, buildings, low low rise, rise, 4 height: 5.1of: m (3.7 m) two multifamilyto four floors, buildings, built from low brickwork rise, 44 height:1 5.1 m (3.72 m) 2 two to four floors, built from brickwork sizeheight:1 : 211.8 5.1 m m (3.7(174.1 m) m ) two totwo four tofloors, four built floors, from built brickwork from size1: 211.8 m² (174.1 m²) or concrete, located in newly developed size1: 211.81 m² (174.1 m²) or concrete, located in newly developed sizelength1: 211.81 :m² 21.8 (174.1 m (8.5 m²) m) or concrete, locatedbrickwork in newly developed length1: 21.8 m (8.5 m) urban areas. length1: 21.81 m (8.5 m) urban areas. lengthwidth11: 21.8: 13.1 m (8.5 m (5.3m) m) or concrete,urban locatedareas. in newly width1: 13.1 m (5.3 m) width1: 13.1 m (5.3 2m) personswidth1/:household 13.1 m (5.32 m): 4.3 (1.6) developed urban areas. persons/household2: 4.3 (1.6) persons/household2: 4.3 (1.6) persons/household2: 4.3 (1.6) Multifamily, local (n = 53) Multifamily, local (n = 53) Multifamily, local (n = 53) Multifamily, local (n = 53) average (standard deviation) Multistory/multiunitMultistory/multiunit apartments apartments with more averageaverage (standard (standard deviation) deviation) Multistory/multiunit of: apartments with more average (standardof: deviation) Multistory/multiunitwiththan more three apartments than units. three with units. more height:of: 5.5 m (4.8 m) than three units. 5 height: 5.5of: m (4.8 m) CCommercialommercialthan and/orthree and units./ orpublic public usage usage 5 height:1 5.5 m (4.82 m) 2 Commercial and/or public usage 55 sizesizeheight:1: 346.2: 346.2 5.5 m² m m(249.1 (4.8(249.1 m) m²) m ) is possible.Commercial Traditional and/or style public of construction usage 1 1 is possible. Traditional style of size1: 346.2 m² (249.1 m²) is possible. Traditional style of construction sizelengthlength: 346.21: 30.3: m² 30.3 m (249.1 (14.1 m (14.1 m²)m) m) is possible. andTraditional local inhabitants. style of construction 1 1 construction and local length1: 30.3 m (14.1 m) and local inhabitants. lengthwidthwidth1:: 30.314.6: 14.6 mm (14.1(7.3 m (7.3m) m) m) and local inhabitants. 1 2 inhabitants. width1: 14.6 m (7.3 m) persons/householdpersonswidth/:household 14.6 m (7.32: 5.0 m): (1.8) 5.0 (1.8) 2 persons/household2: 5.0 (1.8) persons/householdMultifamily, modern: 5.0 (1.8) Multifamily, modern Multifamily,Multifamily,(n = 33) modern modern (n = 33) Multistory/multiunit apartments (n (=n 33)= 33) Multistory/multiunit apartments with more Multistory/multiunitwith more apartme than threents with units. more averageaverage (standard (standard deviation) deviation) Multistory/multiunit of: than three apartme units.nts with more average (standard deviation) Modernthan style threeof units. construction. average (standardof: deviation) Modernthan style three of construction. units. height:of: 8.3 m (10.0 m) Modern style of construction. 66 1 of: 2 2 ModernThis class style only of construction. contains hotels 6 sizeheight:: 8548.3of: m m (10.0(1517.9 m) m ) This classModern only style contains of construction. hotels and other 6 height: 8.3 m (10.0 m) This class only contains hotels and other height:1 8.31 m (10.0 m) This andclass otheronly contains mixed-use hotels commercial and other sizelength1: 854 m²: 43.4 (1517.9 m (28.3 m²) m) mixed-use commercial buildings with non- size1: 854 m² (1517.9 m²) mixed-use commercial buildings with non- size1: 8541 1 m² (1517.9 m²) mixed-usebuildings commercial with buildings non-local with non- lengthwidth1: 43.4: 23.8 m (28.3 m (19.4 m) m) local residents. length1: 43.4 m (28.3 m) local residents. length11: 43.4 m (28.3 m) 3 localresidents. residents. widthpersons1: 23.8/household m (19.4 m) :- width1: 23.8 m (19.4 m) width1: 23.8 m (19.4 3m) persons/household3: - persons/household3: - persons/household3: - Data 2020, 5, 42 6 of 18

Table 1. Cont. Data 2020, 5, x FOR PEER REVIEW 6 of 18 DataID 2020 Building, 5, x FOR TypePEER andREVIEW Statistics Description Representative Reference Picture6 of 18 Special structure / other SpecialSpecial structure(n = structure 11) / other/other (n (=n 11)= 11) averageaverage (standard (standard deviation) deviation) of: average (standard deviation) WideWide range range of built of built-up-up averageheight: (standardof: 5.4 mdeviation) (5.0 m) Wide range of built-up 77 of: structuresstructuresWide range with with aof predominant abuilt predominant-up 7 sizeheight:1: 552.15.4of: m m(5.02 (928m) m2) structures with a predominant 7 height: 5.4 m (5.0 m) structuresnon- residentialwith a predominant use. 1 1 nonnon-residential-residential use. use. sizeheight:length1: 552.1 5.4:39.3 m² m (928(5.0 m (24.0 m)m²) m) size : 552.1 m² (928 m²) non-residential use. 1 1 1 sizelengthwidth: 552.11:39.3: 22.1 m² m (928(24.0 m (14.8 m²) m) m) 11 3 widthlengthpersons1::39.3 22.1/household m (24.0(14.8 m) :- 1 3 widthpersons/household: 22.1 m (14.8 3m): - 3 persons/householdHall (n = 17) : - HallHall (n =(n 17=) 17) average (standard deviation) of: average (standard deviation) Large buildings with one to averageheight: (standardof: 5.1 mdeviation) (3.8 m) Large buildings with one to multiple stores, of:1 2 Large buildingsmultiple with stores, one non-residentialto multiple stores, 8 height:size 5.1of:: m 917.6 (3.8 m) m Largenon buildings-residential with use. one Mostly to multiple markets, stores, 88 height:2 5.1 m (3.81 m) non-residentialuse. Mostly use. Mostly markets, markets, (1654.01 m )length : 44.3 m (31.4 8 sizeheight:1: 917.6 5.1 m² m (1654.0 (3.8 m) m²) nonwarehouses-residentialwarehouses or industrialuse. Mostly or industrial buildings. markets, 1 1 m) sizelength: 917.61: 44.3 m² m (1654.0 (31.4 m)m²) warehouses or buildings.industrial buildings. 11 1 lengthwidthwidth1:: 22.844.3: 22.8 mm (15.8(31.4 m (15.8 m)m) m) 1 3 3 widthpersons/householdpersons: 22.8/household m (15.8 3m): - :- 3 Outbuildingpersons/household / shack (n: =- 16) Outbuilding / shack (n = 16) averageOutbuilding (standard/shack deviation)(n = 16) average (standard deviation) A small, often rundown, non-residential averageaverage (s (standardtandardof: deviation) deviation) of:A small, oftenA small, rundown, often non rundown,-residential of: Abuilding small, often or an rundown, outbuilding non with-residential non- 9 height:height: 2.3of: 2.3m (2.7 m (2.7m) m) buildingnon-residential or an outbuilding building with ornon an- 9 height: 2.3 m (2.7 m) residentialbuilding or usage an outbuilding (e.g., storage, with bicycl non-e 1 residential usage (e.g., storage, bicycle 9 sizeheight:1: 194.0 2.3 m² m (144.8 (2.72 m) m²) 2 9 sizesize: 94.0: 94.0 m² m(144.8(144.8 m²) m ) residentialoutbuilding usageracks). with(e.g., storage, non-residential bicycle 1 1 racks). sizelength: 94.01: 16.81 m² m (144.8 (17.0 m²) m) lengthlength: 16.8: 16.8 m (17.0 m (17.0 m) m) usage (e.g.,racks). storage, bicycle 1 1 1 lengthwidthwidth:1 :16.8 7.9: 7.9mm (4.6(17.0 m (4.6m) m) m) racks). 1 3 3 persons/householdwidthpersons: 7.9/household m (4.6 m)3: - :- 1 persons/household3: - 2 1 as retrieved from the footprints of the surveyed buildings as described in chapter 2.3; 2 as retrieved 1 as retrieved from the footprints of the surveyed buildings as described3. in chapter 2.3; 2 as retrieved 1 asfrom retrieved the data from collection the footprints conducted of the surveyed by Vetter buildings-Gindele as et described al. (2019) in [3] Section; 3. not 2.3 assessed; 2 as retrieved. from the data collectionfrom the conducted data collection by Vetter-Gindele conducted et by al. Vetter (2019)-Gindele [3]; 3. not et assessed. al. (2019) [3]; 3. not assessed. 2.5. External Data 2.5.2.5. External External Data Data To add a spatial relation to the identified changes, administrative boundaries were retrieved from To add a spatial relation to the identified changes, administrative boundaries were retrieved from the To database add aspatial of global relation administrative to the identified areas (GADMchanges,, administrative[11]) which divides boundaries the study were area retrieved into eight from the database of global administrative areas (GADM, [11]) which divides the study area into eight thedistricts database and of51 globalwards, administrative of which 50 belong areas to (GADM, Da Nang [ 11and]) one which belongs divides to the the adjacent study area Quảng into Nam eight districts and 51 wards, of which 50 belong to Da Nang and one belongs to the adjacent Quảng Nam districtsprovince. and We 51 use wards, the wards of which as an 50 administrative belong to Da division Nang and of the one city belongs into homogenous to the adjacent units Quof analysis.ảng Nam province. We use the wards as an administrative division of the city into homogenous units of analysis. province.As a We measure use the which wards is as independent an administrative from political division divisions, of the city hexagon into homogenousal cells withunits a width ofanalysis. of 250 As a measure which is independent from political divisions, hexagonal cells with a width of 250 metersAs a were measure computed. which isCompared independent to squared from political measures, divisions, hexagons hexagonal are more cells suitable with a width to represent of 250 m meters were computed. Compared to squared measures, hexagons are more suitable to represent werenearest computed. neighbor Compared areas, reduce to squaredthe sampling measures, bias at hexagons the edges areand more are reportedly suitable to easier represent to interpret nearest nearest neighbor areas, reduce the sampling bias at the edges and are reportedly easier to interpret neighborby the human areas, eye reduce [12]. the Both sampling measures bias allow at the for edgesa detailed and and are reportedlyprecise analyses easier of to the interpret city without by the violatiby the nghuman the personal eye [12] property. Both measures rights of allow the residents for a detailed or house and owners. precise analyses of the city without humanviolati eyeng the [12 personal]. Both measures property allowrights for of the a detailed residents and or precisehouse owners. analyses of the city without violating violating the personal property rights of the residents or house owners. the personal property rights of the residents or house owners. 2.6. Data Processing and Change Detection 2.6. Data Processing and Change Detection 2.6. DataThe Processing delineation and of Changebuilt-up Detection areas and the extraction of building heights from the Pléiades image of The delineation of built-up areas and the extraction of building heights from the Pléiades image of 2015The was delineation conducted ofusing built-up an object areas-based and image the extraction analysis approach of building, as heightsdescribed from by Warth the Pl éetiades al. [4] image. As described2015 was conducted in that study, using built an object-up areas-based were image identified analysis approachusing a multiresolution, as described by segmentation. Warth et al. [4] This. As ofdescribed 2015 was inconducted that study, using built an-up object-based areas were image identified analysis using approach, a multiresolution as described segmentation. by Warth et This al. [4 ]. described in that study, built-up areas were identified using a multiresolution segmentation. This segmentation was based on the four spectral bands of the acquisition, their three principal components, Assegmentation described in was that based study, on built-up the four areasspectral were bands identified of the acquisition using a multiresolution, their three principal segmentation. components, This a Canny edge detection algorithm [13], and a digital surface model retrieved from the tri-stereoscopic segmentationa Canny edge was detection based onalgorithm the four [13] spectral, and a bands digital of surface the acquisition, model retrieved their three from principal the tri-stereoscopic components, information of the Pléiades product [14]. The image objects were then classified into “built-up” and a Cannyinformation edge of detection the Pléiades algorithm product [13 [14]], and. The a digital image surfaceobjects modelwere then retrieved classified from into the “built tri-stereoscopic-up” and “non-built-up” using rule-based thresholds applied to the derived object features, resulting in an information“non-built-up” of the using Pléiades rule-based product thresholds [14]. The applied image objects to the derivedwere then object classified features, into resulting “built-up” in an and accuracy of 88.4%. All built-up areas were divided into single buildings using spatial intersection with “non-built-up”accuracy of 88.4% using. All rule-based built-up areas thresholds were divided applied into tosingle the derivedbuildings object using features,spatial intersection resulting with in an cadastral boundaries and visual identification, as described by Vetter-Gindele et al. (2019) [3]. The accuracycadastral of boundaries 88.4%. All built-up and visual areas identification were divided, as into described single by buildings Vetter-Gindele using spatial et al. intersection (2019) [3]. The with geometric accuracy of single buildings was assessed by the manual digitization of 50 randomly selected cadastralgeometric boundaries accuracy of and single visual buildings identification, was assessed as describedby the manual by Vetter-Gindeledigitization of 50 et randomly al. (2019) selected [3]. The buildings, whose areas were then compared to the automatically derived building footprints. The geometricbuildings accuracy, whose ofarea singles were buildings then compared was assessed to the by automatically the manual digitization derived building of 50 randomly footprints. selected The comparison resulted in an overall agreement of 87.2%. While accuracies of over 95% were achieved for comparison resulted in an overall agreement of 87.2%. While accuracies of over 95% were achieved for2 buildings of average sizes (50–160 m²), larger underestimations were observed for very small (<50 m2) buildings of average sizes2 (50–160 m²), larger underestimations were observed for very small (<50 m2) and very large (>200 m2) buildings, which only make up a small fraction of the buildings in Da Nang. and very large (>200 m2) buildings, which only make up a small fraction of the buildings in Da Nang. Data 2020, 5, 42 7 of 18 buildings, whose areas were then compared to the automatically derived building footprints. The comparison resulted in an overall agreement of 87.2%. While accuracies of over 95% were achieved for buildings of average sizes (50–160 m2), larger underestimations were observed for very small (<50 m2) and very large (>200 m2) buildings, which only make up a small fraction of the buildings in Da Nang. DataInData the 20 2020 end,20,, 5,,, 5xx, theFORxFOR FOR derived PEERPEER PEER REVIEWREVIEW REVIEW footprints resulted in building areas which were found to be reliable enough7 7of of 18 18 to serve as one of the variables for the classification of building types in the next step. InIn the the end, end, the the derived derived footprints footprints resulted resulted in in building building areas areas which which were were found found to to be be reliable reliable enough enough to to InIn the theAccordingly, end, end, the the derived derived building footprints footprints types resulted resulted were assignedin in building building based areas areas on which which the criteriawere were found found given to to be in be reliable Table reliable1 andenough enough manual to to serveserve as as one one of of the the variables variables for for the the classification classification of of building building types types in in the the next next step. step. serverefinement.serve as as one one of Thisof the the classificationvariables variables for for the typologythe classification classification resulted of of building inbuilding an overall types types in number in the the next next of step. 241,217step. building features AccordinglyAccordingly, building, building types types were were assigned assigned based based on on the the criteri criteria agiven given in in Table Table 1 1and and manual manual in theAccordinglyAccordingly study area, forbuilding, building the year types types 2015. were were To assigned understandassigned based based the on qualityon the the criteri criteri of thea agiven buildinggiven in in Table typeTable 1 classification, 1and and manual manual an refinement.refinement. This This classification classification typology typology resulted resulted in in an an overall overall number number of of 241,217 241,217 building building feature features ins in refinement.accuracyrefinement. assessment This This classification classification was conducted typology typology based resulted resulted on the in in 803an an overall building overall number number references of of 241,217 241,217 which building werebuilding collected feature feature durings sin in thethe study study area area for for the the year year 2015 2015.. To. To unde understandrstand the the quality quality of of the the building building type type classification, classification, an an thethe field study campaign, area for the as described year 2015 in. To Section under 2.3stand[ 3]. the The quality confusion of the matrix building between type classification, the reference dataan accuracyaccuracy assessment assessment was was conducted conducted based based on on the the 803 803 building building references references which which were were collected collected and the classified building types is shown in Table A1 in the AppendixA and discussed in more detail duringduring the the field field campaign campaign,, as, as described described in in section section 2.3. 2.3. [3] [3].. The. The confusion confusion matrix matrix between between the the reference reference in Section 3.5. An overall classification accuracy of 90.9% was achieved. datadata and and the the classified classified building building types types is is shown shown in in Table Table A1 A1 in in the the appendix appendix and and d iscusseddiscussed in in more more In the next step, a typology of changes was developed, consisting of five classes (Table2): no detaildetail in in section section 3.5. 3.5. An An overall overall classification classification accuracy accuracy of of 90.9% 90.9% was was achieved. achieved. change (1), construction of a new building (2), changes in the building, e.g., roof color or size, while InIn the the next next ste step,p, a atypology typology of of changes changes was was developed, developed, consisting consisting of of five five classes classes (Table (Table 2): 2): no no maintaining the building type according to Table1 (3), upgrading of the building to a new building changechange (1), (1), construction construction of of a anew new building building (2), (2), changes changes in in the the building, building, e.g. e.g.,, roof, roof color color or or size, size, while while type, mostly of higher quality (4), and demolition of the building (5). All buildings identified for the maintainingmaintaining the the building building type type according according to to Table Table 1 1(3), (3), upgrading upgrading of of the the building building to to a anew new building building year 2015 were then examined for changes by a visual comparison of the two images and one type of type,type, mostly mostly of of higher higher quality quality (4), (4), and and demolition demolition of of the the building building (5). (5). All All buildings buildings identified identified for for the the yearchangeyear 201 201 was5 5were were assigned then then examined examined to each feature.for for changes changes In the by by casea avisual visual of a comparison newlycomparison constructed of of the the two two building, images images and a and new one one feature type type of wasof changemanuallychange was was digitized assigned assigned at to thisto each each location. feature. feature. This In In the resultedthe case case of inof a a anewly totalnewly number constructed constructed of 260,608 building, building, buildings a anew new forfeature feature 2017, was whichwas manuallyindicatesmanually andigitized digitized increase at at ofthis this 19,391 location. location. for the This This investigated resulted resulted in area.in a atotal total numb numberer of of 260,608 260,608 buildings buildings for for 2017 2017,, , whichwhichThe indicates indicates process an an of increase increase visual of interpretation of 19,391 19,391 for for the the was investigated investigated conducted area. area. by independent persons, with additional qualityTheThe control process processby of of visual another visual interpretation interpretation person to ensure was was conducted conducted the highest by by independent level independent of objectivity persons persons, and, with, with to additional minimizeadditional quality errorsquality of controlomissioncontrol by by another and another commission. person person to to ensure However,ensure the the highest highest as demonstrated level level of of objectivity objectivity by Figure and and to3 ,to minimize the minimize identification errors errors of of omission of omission changes and and can commission.stillcommission. be subject However However to the,, interpreter’s asas, as demonstrateddemonstrated demonstrated perception byby by Figure Figure of 3 the ,,3 thethe, the two identificationidentification identification images and ofof oftheir changeschanges changes ability cancan can tostillstill still identify bebe be subjectsubject subject changes. toto to theThisthe interpreter’s interpreter’s was especially perception perception challenging of of the the two intwo denselyimages images and built-upand their their ability areasability to whereto identify identify buildings changes. changes. stand This This was extremelywas especi especiallyally close challenging(Figurechallenging4, left in in part).densely densely This built built required-up-up areas areas a more where where time-consuming buildings buildings stand stand inspectionextremely extremely close and close an (Figure (Figure independent 4 ,,4 leftleft, left part)part) part) crosscheck.. This. This requiredinrequired those areas.a amore more For time time the-consuming-consuming more regular inspection inspection parts ofand and Da an Nangan independent independent (Figure4 crosscheck, rightcrosscheck part) in changesin those those areas areas were.. ForFor. clearlyFor thethe the moremore visiblemore regularandregular unambiguous. parts parts of of Da Da Nang Nang (Figure (Figure 4 ,,4 rightright, right partpart part) changes) changes were were clearly clearly visible visible and and unambiguous. unambiguous. TheTheThe selection selection selection of of a a a manual manual manual assignment assignment assignment of of building of building building types types types for for 2017 for 2017 2017and and the and the derivation derivation the derivation of of the the of correspondingthecorresponding corresponding changes changes changes was was was considerably considerably considerably faster faster faster than than than repeating repeating repeating the the the semi semi semi-automated-automated-automated building building building classificationclassificationclassification as as itit it waswas was conductedcond conducteducted for for for the the the image image image of of 2015.of 2015. 2015. At At theAt the the same same same time, time, time, it allowedit itallowed allowed us tous us achieve to to achieve achieve higher higheraccuracieshigher accuracies accuracies in the final inin data thethe product—for final final data data product example, product——for byfor manually example example,, correcting, byby manually manually smaller correcting correcting misclassifications smaller smaller of misclassifications of the building types of 2015 while searching for changes between both images. misclassificationsthemisclassifications building types of of the the 2015 building building while types searching types of of 2015 2015 for while changeswhile searching searching between for for both changes changes images. between between both both images images. .

TableTable 2. 2. Definition Definition of of changes changes and and examples examples for for the the two two satellite satellite images images used used in in this this study study. . TableTableTable 2. 2. 2. DefinitionDeﬁnition Definition of ofof changes changes and and examples examples for for for the the the two two two satellite satellite satellite images images images used used used in in this in this this study study study.. . IDID ChangeChange ExampleExample 2015 2015 ExampleExample 2017 2017 IDID ID ChangeChange Change Example ExampleExample 20152015 2015 ExampleExample Example 2017 20172017

1 1 NoNo ch changeange 1 1 1NoNo No ch ch changeangeange

2 2 2NewNew New b uildingb buildinguilding

ChangeChange/re/renovationnovation withoutwithout a asignificant significant 3 3 changechange in in the the building building typetype or or footprint footprint DataData 20 202020, 5,, 5x, FORx FOR PEER PEER REVIEW REVIEW 7 7of of 18 18

InIn the the end, end, the the derived derived footprints footprints resulted resulted in in building building areas areas which which were were found found to to be be reliable reliable enough enough to to serveserve as as one one of of the the variables variables for for the the classification classification of of building building types types in in the the next next step. step. AccordinglyAccordingly, building, building types types were were assigned assigned based based on on the the criteri criteria agiven given in in Table Table 1 1and and manual manual refinement.refinement. This This classification classification typology typology resulted resulted in in an an overall overall number number of of 241,217 241,217 building building feature features ins in thethe study study area area for for the the year year 2015 2015. To. To unde understandrstand the the quality quality of of the the building building type type classification, classification, an an accuracyaccuracy assessment assessment was was conducted conducted based based on on the the 803 803 building building references references which which were were collected collected duringduring the the field field campaign campaign, as, as described described in in section section 2.3. 2.3. [3] [3]. The. The confusion confusion matrix matrix between between the the reference reference datadata and and the the classified classified building building types types is is shown shown in in Table Table A1 A1 in in the the appendix appendix and and d iscusseddiscussed in in more more detaildetail in in section section 3.5. 3.5. An An overall overall classification classification accuracy accuracy of of 90.9% 90.9% was was achieved. achieved. InIn the the next next ste step,p, a atypology typology of of changes changes was was developed, developed, consisting consisting of of five five classes classes (Table (Table 2): 2): no no changechange (1), (1), construction construction of of a anew new building building (2), (2), changes changes in in the the building, building, e.g. e.g., roof, roof color color or or size, size, while while maintainingmaintaining the the building building type type according according to to Table Table 1 1(3), (3), upgrading upgrading of of the the building building to to a anew new building building type,type, mostly mostly of of higher higher quality quality (4), (4), and and demolition demolition of of the the building building (5). (5). All All buildings buildings identified identified for for the the yearyear 201 2015 5were were then then examined examined for for changes changes by by a avisual visual comparison comparison of of the the two two images images and and one one type type of of changechange was was assigned assigned to to each each feature. feature. In In the the case case of of a anewly newly constructed constructed building, building, a anew new feature feature was was manuallymanually digitized digitized at at this this location. location. This This resulted resulted in in a atotal total numb numberer of of 260,608 260,608 buildings buildings for for 2017 2017, , whichwhich indicates indicates an an increase increase of of 19,391 19,391 for for the the investigated investigated area. area. TheThe process process of of visual visual interpretation interpretation was was conducted conducted by by independent independent persons persons, with, with additional additional quality quality controlcontrol by by another another person person to to ensure ensure the the highest highest level level of of objectivity objectivity and and to to minimize minimize errors errors of of omission omission and and commission.commission. However However, as, as demonstrated demonstrated by by Figure Figure 3 ,3 the, the identification identification of of changes changes can can still still be be subject subject to to thethe interpreter’s interpreter’s perception perception of of the the two two images images and and their their ability ability to to identify identify changes. changes. This This was was especi especiallyally challengingchallenging in in densely densely built built-up-up areas areas where where buildings buildings stand stand extremely extremely close close (Figure (Figure 4 ,4 left, left part) part). This. This requiredrequired a amore more time time-consuming-consuming inspection inspection and and an an independent independent crosscheck crosscheck in in those those areas areas. For. For the the more more regularregular parts parts of of Da Da Nang Nang (Figure (Figure 4 ,4 right, right part part) changes) changes were were clearly clearly visible visible and and unambiguous. unambiguous. TheThe selection selection of of a a manual manual assignment assignment of of building building types types for for 2017 2017 and and the the derivation derivation of of the the correspondingcorresponding changes changes was was considerably considerably faster faster than than repeating repeating the the semi semi-automated-automated building building classificationclassification as as it itwas was cond conducteducted for for the the image image of of 2015. 2015. At At the the same same time, time, it itallowed allowed us us to to achieve achieve higherhigher accuracies accuracies inin thethe final final data data product product——forfor example example, , byby manually manually correcting correcting smaller smaller misclassificationsmisclassifications of of the the building building types types of of 2015 2015 while while searching searching for for changes changes between between both both images images. .

TableTable 2. 2.Definition Definition of of changes changes and and examples examples for for the the two two satellite satellite images images used used in in this this study study. .

IDID ChangeChange ExampleExample 2015 2015 ExampleExample 2017 2017

1 1 NoNo ch changeange

Data 2020, 5, 42 8 of 18 2 2 NewNew b uildingbuilding Table 2. Cont.

ID Change Example 2015 Example 2017

ChangeChange/re/renovationnovation Change/renovation withoutwithoutwithout a asignificant signiﬁcantsignificant 3 3 3 changechangechange in in in the the building buildingbuilding DataDataData 20 20 20202020, 5, ,,5 5x, , x FORx FOR FOR PEER PEERPEER REVIEW REVIEWREVIEW 8 8of of 818 of18 18 DataData 20 202020, 5,, 5xtype, FORxtypetype FOR or PEER or orPEER footprint footprint REVIEW REVIEW 8 8of of 18 18

UpgradingUpgradingUpgrading including includingincluding including a aa 4 44 significantsignificantsignificanta signiﬁcant change change change in inin the in thethe 4 4 4 significantsignificant change change in in the the buildingbuildingthe type type building/footprint//footprintfootprint type/footprint

DemolitionDemolitionDemolition o of foo theff the thethe 5 55 5 buildingbuildingbuilding

InIn the the final final step, step, the the building building types types and and their their changes changes between between 2015 2015 and and 2017 2017 were were summarized summarized InInInIn the the thethe final ﬁnal finalfinal step, step, step,step, the thethe building buildingbuilding types typestypes and andand their their theirtheir changes changes changeschanges between between betweenbetween 2015 2015 20152015 and and andand 2017 2017 20172017 were were werewere summarized summarized summarizedsummarized forfor each each ward ward (Figure (Figure 4 ,4 black, black dashed dashed lines) lines) a ndand hexagon hexagon (Figure (Figure 4 ,4 blue, blue lines). lines). As As described described in in Table Table forforforfor each eacheach each ward ward wardward ( (Figure Figure ((FigureFigure 4 ,,4 blackblack, black dashed dasheddashed lines) lines)lines) a and nda andnd hexagon hexagon hexagon ( (FigureFigure ( Figure(Figure 44, ,4 blueblue,4 blue, blue lines).lines). lines). lines). As As As Asdescribed described described described in in in Table Table inTable Table 3, 3,3 each, each polygon polygon contains contains attributes attributes on on the the absolute absolute number number of of buildings buildings for for both both years, years, as as well well as as the the 3each3,3 ,each ,each each polygon polygon polygonpolygon contains contains containscontains attributes attributes attributes onon onon the thethe absolute absoluteabsolute number number number number of ofof buildingsof buildingsbuildings buildings for for for forboth both both both years, years, years, years, as as aswell as well wellwell as as the as asthe thethe absoluteabsolute and and relative relative amount amount of of change. change. absoluteabsoluteabsolute and and andand relative relative relativerelative amount amountamount of ofof change. change. change.change.

FigureFigure 4. 4. Example Example for for Hòa Hòa Minh Minh ward. ward. Top: Top: satellite satellite im imimageage from from 2017; 2017; bottom bottom: result:: result of of the the visual visual identificationFigureFigureidentificationidentification 4.4. ExampleExample of ofof changes changeschanges forfor H,Hòa includingò,, aincludingincluding MinhMinh ward.theward. thethe boundaries boundariesboundaries Top:Top: satellitesatellite of ofof administrative administrativeadministrative imageimage fromfrom wards 2017;2017; wardswards bottom:andbottom andand hexagons. hexagons.hexagons.: resultresult ofof thethe visualvisual identiﬁcationidentification ofof changes,changes, including including the the boundaries boundaries of of administrative administrative wards wards and and hexagons. hexagons. 3.3.DataData Description Description 3.3. DataData DescriptionDescription 3.1.3.1. Datatype Datatype 3.1.3.1. Datatype Datatype 3.1. Datatype TheThe two two datasets datasets presented presented in in this this descriptor descriptor (please (please see see the the zip zip file file in in the the Supplement Supplementaryary The two datasets presented in this descriptor (please see the zip ﬁle in the Supplementary Materials) Materials)Materials)The twowill will bedatasets be described described presented in in the the next innext thissections sections descriptor and and the the (please changes changes see in in the the building zip building file instock stock the for Supplementaryfor the the urban urban will be described in the next sections and the changes in the building stock for the urban areas of Da areasMaterials)areas of of Da Da willNang Nang be city, describedcity, aggregated aggregated in the by bynext administrative administrative sections and wards wardsthe changes and and a asystematic systematicin the building grid grid of ofstock hexagon hexagon for theof of 250 urban250 Nang city, aggregated by administrative wards and a systematic grid of hexagon of 250 m in width, meterareasmeters of sin in Dawidth width Nang, will,, willwillcity, be bebe aggregatedexplained explainedexplained. Both..by BothBoth administrative are areare stored storedstored as as vectorwards vector datasetsand datasets a systematic in in the the ESRI ESRI grid shapefile shapefile of hexagon format format of, 250,, whichmwhich in width is is compatible compatible, will be withexplained with common common. Both Geographic Geographic are stored Information Informationas vector data Systems Systemssets in (GIS), (GIS),the ESRI such such shapefileas as QGIS QGIS or format or ArcMap. ArcMap., which EachisEach compatible dataset dataset consists consistswith common of of several several Geographic separate separate files filesInformation with with the the same Systemssame name, name, (GIS), but but differentsuch different as QGISextensions: extensions: or ArcMap. a avector vector Each geometrygeometry (*.shp), (*.shp), an an attrib attributeute table table (*.dbf), (*.dbf), a aprojection projection file file (*.prj), (*.prj), and and an an index index linking linking the the attribute attribute geometrydatasetgeometry consists (*.shp), (*.shp), an of an attribseveral attributeu te separatetable table (*.dbf), (*.dbf), files a withaprojection projection the same file file (*.prj), (*.prj), name, and and but an an different index index linking linking extensions: the the attribute attribute a vector tabletable to to the the geometries geometries (*.shx). (*.shx). Both Both datasets datasets are are stored stored in in the the UTM UTM coordinate coordinate system, system, zone zone 49N 49N tablegeometrytable to to the the (*.shp), geometries geometries an attrib (*.shx). (*.shx).ute tableBoth Both (datasets *.dbf),datasets a are projectionare stored stored infile in the the(*.prj), UTM UTM and coordinate coordinate an index system, linkingsystem, zone thezone attribute 49N 49N (EPSG:(EPSG: 32649) 32649). The. The encoding encoding of of the the attribute attribute table table (*.cpg) (*.cpg) is is UTF UTF-8- 8to to support support the the correct correct display display of of (EPSG:table(EPSG: to 32649) 32649)the geometries. The. The encoding encoding (*.shx). of of the Boththe attribute attribute datasets table table are (*.cpg) (*.cpg)stored is is inUTF UTF the-8 - UTM8to to support support coordinate the the correct correct system, display display zone of 49Nof VietnameseVietnamese characters. characters. Vietnamese(EPSG:Vietnamese 32649) characters. characters.. The encoding of the attribute table (*.cpg) is UTF-8 to support the correct display of Vietnamese characters.

Data 2020, 5, 42 9 of 18 will be explained. Both are stored as vector datasets in the ESRI shapefile format, which is compatible with common Geographic Information Systems (GIS), such as QGIS or ArcMap. Each dataset consists of several separate files with the same name, but different extensions: a vector geometry (*.shp), an attribute table (*.dbf), a projection file (*.prj), and an index linking the attribute table to the geometries (*.shx). Both datasets are stored in the UTM coordinate system, zone 49N (EPSG: 32649). The encoding of the attribute table (*.cpg) is UTF-8 to support the correct display of Vietnamese characters.

3.2. Data Structure Table3 summarizes the attributes contained by the administrative (adm) and hexagonal (hex) datasets. Both datasets contain the absolute number of changes (c1 to c5, as defined in Table2), the absolute number of buildings per year and type (t2015_1 to t2017_9, as defined in Table1), as well as the absolute (nc) number of changes per building type in each ward and hexagon. The relative number of changes (pc) denotes the share of the changes in one building type compared to the total number of changes in the reference area. For the administrative dataset, the results can also be summarized by district, using the respective attributes (dist_ID or district). As one suggestion for further spatial analyses, the hexagon dataset additionally contains an attribute named dist_mean, which defines the Euclidean distance of each hexagon from the historic city center in Hải Châu district. This will be discussed in Section 3.4.

Table 3. Attributes of the two shapeﬁles.

Attribute Description adm hex Source province name of the province x GADM dist_ID id of the district x GADM district name of the district x GADM ward_ID id of the ward x GADM ward name of the ward x GADM ward_en name of the ward without special characters x GADM area area of the ward in square kilometers x GIS analysis per_cover % of ward area covered by satellite x GIS analysis hex_ID id of the hexagon x GIS analysis dist_mean mean distance from the historic city center x GIS analysis c1 no of buildings with no-change x x Satellite image c2 no. of new buildings x x Satellite image c3 no. of changed buildings (same type) x x Satellite image c4 no. of upgraded buildings (other type) x x Satellite image c5 no. of demolished buildings x x Satellite image c_sum no. of all changed buildings x x c2 + c3 + c4 + c5 t2015_1 no. of “basic” buildings in 2015 x x Satellite image t2015_2 no. of “local-type shophouse” buildings in 2015 x x Satellite image t2015_3 no. of “bungalow” buildings in 2015 x x Satellite image t2015_4 no. of “villa-type” buildings in 2015 x x Satellite image t2015_5 no. of “apartment, local” buildings in 2015 x x Satellite image t2015_6 no. of “apartment, modern” buildings in 2015 x x Satellite image t2015_7 no. of “hall” buildings in 2015 x x Satellite image t2015_8 no. of “outbuilding” buildings in 2015 x x Satellite image t2015_9 no. of “special” buildings in 2015 x x Satellite image t2015_sum no. of all buildings in 2015 x x sum(t2015) t2017_1 no. of “basic” buildings in 2017 x x Satellite image t2017_2 no. of “local-type shophouse” buildings in 2017 x x Satellite image t2017_3 no. of “bungalow” buildings in 2017 x x Satellite image t2017_4 no. of “villa-type” buildings in 2017 x x Satellite image t2017_5 no. of “apartment, local” buildings in 2017 x x Satellite image t2017_6 no. of “apartment, modern” buildings in 2017 x x Satellite image t2017_7 no. of “hall” buildings in 2017 x x Satellite image t2017_8 no. of “outbuilding” buildings in 2017 x x Satellite image t2017_9 no. of “special” buildings in 2017 x x Satellite image t2017_sum no. of all buildings in 2017 x x sum(t2017) Data 2020, 5, 42 10 of 18

Table 3. Cont.

Attribute Description adm hex Source nc_t1 net change in “basic” buildings x x t2017_1-t2015_1 nc_t2 net change in “local-type shophouse” buildings x x t2017_2-t2015_2 nc_t3 net change in “bungalow” buildings x x t2017_3-t2015_3 nc_t4 net change in “villa-type” buildings x x t2017_4-t2015_4 nc_t5 net change in “apartment, local” buildings x x t2017_5-t2015_5 nc_t6 net change in “apartment, modern” buildings x x t2017_6-t2015_6 nc_t7 net change in “hall” buildings x x t2017_7-t2015_7 nc_t8 net change in “outbuilding” buildings x x t2017_8-t2015_8 nc_t9 net change in “special” buildings x x t2017_9-t2015_9 nc_all net change in all buildings x x t2017_sum-t2015_sum pc_t1 % of change in “basic” buildings x x nc_t1/t2015_1 pc_t2 % of change in “local-type shophouse” buildings x x nc_t2/t2015_2 pc_t3 % of change in “bungalow” buildings x x nc_t3/t2015_3 pc_t4 % of change in “villa-type” buildings x x nc_t4/t2015_4 pc_t5 % of change in “apartment, local” buildings x x nc_t5/t2015_5 pc_t6 % of change in “apartment, modern” buildings x x nc_t6/t2015_6 pc_t7 % of change in “hall” buildings x x nc_t7/t2015_7 pc_t8 % of change in “outbuilding” buildings x x nc_t8/t2015_8 pc_t9 % of change in “special” buildings x x nc_t9/t2015_9 pc_all % of changed buildings based on all buildings in 2015 x x c_sum/t2015_sum

Accordingly, the dataset does not contain any data related to individuals collected during the ﬁeld campaigns. The attributes are purely a result of the geospatial analyses described in the methodology section. This aggregation grants that no user can calculate the original building or plot geometries, nor their exact type or location.

3.3. Administrative Level The dataset contains the changes in the building stock aggregated by wards. These represent the official administrative units, so this dataset is best compatible with other socio-demographic data, such as the official Vietnam Population and Housing Census released by the General Statistics Office of Vietnam [15]. However, it must be kept in mind that, in particular, the wards at the edges are not entirely covered by the satellite images, so absolute numbers have to be interpreted with care. For this reason, the attribute table of the administrative dataset contains an additional attribute, “per_cover”, which tells the percentage of each ward’s area inside the AOI. It is illustrated in Figure A1 in the AppendixB, which highlights the ten wards which are only covered by 70% or less of the satellite images. Figure5 shows the dominance of new construction activities over demolitions in the observed time between 2015 and 2017. Except for the already very densely built-up wards of An Hải Đông, Bình Hiên, and Phước Ninh, which are located near the city center, and the rural wards of Hòa Vang district in the western outskirts of the city, all parts of Da Nang have experienced a considerable increase in building numbers, mostly exceeding 500 new buildings per ward. In contrast, the number of demolitions is visibly lower throughout the city. Over 90% of all wards range between 10 and 200 demolished buildings. Of course, these numbers are strongly dependent on the size of a ward. For this reason, the attribute “area” is contained in the dataset to calculate the absolute changes per square kilometer (c1/area or c5/area), as shown in Figure6, which gives a more representative picture for each ward. It highlights the increase in Hòa Minh in the north and the relative high demolitions in Hải Châu district, representing the historic city center with its extreme building density. Another method of calculating a spatially uncorrelated representation of changes is given by the hexagons described in Section 3.4. DataData2020 2020,,5 5,, 42 x FOR PEER REVIEW 1111 of 1818 Data 2020, 5, x FOR PEER REVIEW 11 of 18

Figure 5. Absolute building changes between 2015 and 2017: new buildings (c2, left) and demolitions FigureFigure 5.5. Absolute building changes between 2015 and 2017: new buildings ( (c2,c2, left)left) andand demolitionsdemolitions (c5, right). (c5,(c5, right). right).

FigureFigure 6. 6.Changes Changes per per square square kilometer kilometer between between 2015 and 2015 2017: and new 2017: buildings new buildings (c2, left) and (c2 demolitions, left) and Figure 6. Changes per square kilometer between 2015 and 2017: new buildings (c2, left) and (c5,demolitions right). (c5, right). demolitions (c5, right). Figure7 demonstrates further spatial patterns revealed in the data: It shows that multifamily Figure 7 demonstrates further spatial patterns revealed in the data: It shows that multifamily apartmentsFigure largely7 demonstrates increased further between spatial 2015 patterns and 2017, revealed but with in the spatially data: It distinct shows patterns.that multifamily While apartments largely increased between 2015 and 2017, but with spatially distinct patterns. While local- local-typeapartments apartments largely increased (building between type 5 in2015 Table and2), 2017, which but are with predominantly spatially distinct inhabited patterns. by local While people, local- type apartments (building type 5 in Table 2), which are predominantly inhabited by local people, showtype apar indiﬀtmentserentpatterns (building and type even 5 in slight Table decreases 2), which in theareouter predominantly wards, there inhabited is a tremendous by localincrease people, show indifferent patterns and even slight decreases in the outer wards, there is a tremendous increase inshow hotels indifferent (modern-type pattern apartments,s and even slight building decreases type 6 inin Tablethe outer2) along wards, the there coasts is in a tremendous the east (Ng increaseũ Hành in hotels (modern-type apartments, building type 6 in Table 2) along the coasts in the east (Ngũ Hành Sinơ nhotels district) (modern and in-type the northapartments, (Liên Chi buildingểu district), type 6 which in Table underlines 2) along Dathe Nang’scoasts in ambitions the east (Ngũ to increase Hành Sơn district) and in the north (Liên Chiểu district), which underlines Da Nang’s ambitions to increase itsSơn touristic district) sector and in [16 the]. north (Liên Chiểu district), which underlines Da Nang’s ambitions to increase its touristic sector [16]. its touristicHowever, sector asalready [16]. stated above, the size of the wards largely diﬀers and ranges between 0.23 and 47 square kilometers, with an average ward size of 7.1 square kilometers. Accordingly, absolute building numbers and their changes are spatially biased by the ward size. For this reason, a second dataset is introduced, which provides a regular representation of the aggregated results.

Data 2020, 5, x FOR PEER REVIEW 12 of 18

Data 2020, 5, 42 12 of 18 Data 2020, 5, x FOR PEER REVIEW 12 of 18

Figure 7. Relative increase in local (pc_t5, left) and modern (pc_t6, right) apartments between 2015 and 2017 per ward.

However, as already stated above, the size of the wards largely differs and ranges between 0.23 and 47 square kilometers, with an average ward size of 7.1 square kilometers. Accordingly, absolute building numbers and their changes are spatially biased by the ward size. For this reason, a second

dataset is introduced, which provides a regular representation of the aggregated results. FigureFigure 7.7. RelativeRelative increaseincrease ininlocal local(pc_t5, (pc_t5 left), left) and and modern modern (pc_t6, (pc_t6 right), right) apartments apartments between between 2015 2015 and 3.4. Spatial2017and 20 per 1Aggregation7 ward. per ward. 3.4. Spatial Aggregation TheHowever, dataset as < alreadyDaNang_buildings_hex250m_UTM4 stated above, the size of the wards9N.shp largely> aggregates differs theand changes ranges betweenat a spatially 0.23 regularand The47 squaregrid, dataset where kilometers enclosesaggregates kilometers. an thearea Accordingly, changes of 72 hectares. at a spatiallyabsolute This regularallowsbuilding us grid, numbers to identify where and each patterns their hexagon changes that has are anotare width relatedspatially of 250 to biasedadministrative m and by encloses the ward boundaries an areasize. of For 72. As hectares.this an reason, example, This a second allowsFigure us8dataset shows to identify is the introduced absolute patterns ,that differenceswhich are provides not relatedin bungalows a regular to administrative andrepresentation villa-type boundaries. buildingsof the aggregated As calculated an example, results. by Figure the following8 shows theequations: absolute t2017_3 diﬀerences - int2015_3 bungalows and and t2017_4 villa-type - t2015_4 buildings. calculatedIt shows that by these the following building equations:types and t2017_3-t2015_3their3.4. Spatial changes Aggregation are and not t2017_4-t2015_4. drastically shaping It shows the morphology that these buildingof the city, types as most and theirof the changes hexagons are have not drasticallyvalues close shaping to zero. theHowever, morphology it also ofshows the city,that asmost most of the of thebungalows hexagons were have demolished values close in the to zero.new The dataset aggregates the changes at a spatially However,urban development it also shows areas that of most Hòa ofXuân the bungalows(Cẩm Lệ district) were demolished and Hòa Quy in the (Ngũ new Hành urban Sơn development district) in regular grid, where each hexagon has a width of 250 m and encloses an area of 72 hectares. This areasthe southwest of Hòa Xu ofân the (C ẩcitym L centerệ district), which and is H oneòa Quy of Da (Ng Nangũ Hà’snh fastest Sơn district) urbanizing in the areas. southwest Large ofmasses the city of allows us to identify patterns that are not related to administrative boundaries. As an example, Figure center,sand were which piled is one to create of Da Nang’snew space fastest for development urbanizing areas. at a confluence Large masses area of of sand the wereHán piledRiver. to In create turn, 8 shows the absolute differences in bungalows and villa-type buildings calculated by the following newmost space of the for new development villa-type buildings at a conﬂuence are constructed area of the along Hán River.the coastal In turn, area most in the of east the newto increa villa-typese the equations: t2017_3 - t2015_3 and t2017_4 - t2015_4. It shows that these building types and buildingsnumber of are holiday constructed residences. along the coastal area in the east to increase the number of holiday residences. their changes are not drastically shaping the morphology of the city, as most of the hexagons have values close to zero. However, it also shows that most of the bungalows were demolished in the new urban development areas of Hòa Xuân (Cẩm Lệ district) and Hòa Quy (Ngũ Hành Sơn district) in the southwest of the city center, which is one of Da Nang’s fastest urbanizing areas. Large masses of sand were piled to create new space for development at a confluence area of the Hán River. In turn, most of the new villa-type buildings are constructed along the coastal area in the east to increase the number of holiday residences.

FigureFigure 8. AbsoluteAbsolute change change in bungalow in bungalow-type-type (left) (left)and villa and-type villa-type (right) (right)buildings buildings between between2015 and 20152017. and 2017.

The notable increase in buildings in Cam Le Cẩm Lệ is furthermore underlined by Figure9, where two hotspots of new buildings (change type 2 as deﬁned in Table1) are visible: these are the wards of An Khê,Hòa Phát, and Hòa Minh in the north and the new construction areas in Hòa Xuân ward of Cẩm Lệ district enclosed by the Hàn River in the southwest. In particular, the increase in built-up areasFigure in Cẩ m8. Absolute Lệ was alreadychange in conﬁrmed bungalow-type in 2016 (left) by and Bachofer villa-type and (right) Rau buildings [17] and between Linh and2015 Chuongand 2017. [18].

Data 2020, 5, x FOR PEER REVIEW 13 of 18 Data 2020, 5, x FOR PEER REVIEW 13 of 18 The notable increase in buildings in Cam Le Cẩm Lệ is furthermore underlined by Figure 9, The notable increase in buildings in Cam Le Cẩm Lệ is furthermore underlined by Figure 9, where two hotspots of new buildings (change type 2 as defined in Table 1) are visible: these are the where two hotspots of new buildings (change type 2 as defined in Table 1) are visible: these are the wards of An Khê, Hòa Phát, and Hòa Minh in the north and the new construction areas in Hòa Xuân wards of An Khê, Hòa Phát, and Hòa Minh in the north and the new construction areas in Hòa Xuân Dataward2020 of, 5 Cẩm, 42 Lệ district enclosed by the Hàn River in the southwest. In particular, the increase13 of 18in ward of Cẩm Lệ district enclosed by the Hàn River in the southwest. In particular, the increase in built-up areas in Cẩm Lệ was already confirmed in 2016 by Bachofer and Rau [17] and Linh and built-up areas in Cẩm Lệ was already confirmed in 2016 by Bachofer and Rau [17] and Linh and Chuong [18]. This is especially alarming because these new residential areas are expected to lie below ThisChuong is especially [18]. This alarming is especially because alarming these newbecause residential these new areas residential are expected areas to are lie expected below the to tideline lie below in the tideline in 2050 [19] (Figure A3). In turn, most of the building changes (within the same type, 2050the tideline [19] (Figure in 20 A350 ).[19] In turn,(Figure most A3). of In the turn, building most changes of the building (within thechanges same (within type, change the same type type, 3 as change type 3 as defined in Table 1) occurred in the districts Hải Châu and Ngũ Hành Sơn, around deﬁnedchange intype Table 3 as1) defined occurred in in Table the districts 1) occurred H ải Chin âtheu and districts Ngũ HHảiành Châu Sơn, and around Ngũ the Hành historic Sơn center, around of the historic center of Da Nang. These are the parts of the city which are already extremely dense DatheNang. historic These center are of the Da parts Nang of.the These city are which the areparts already of the extremely city which dense are (Figurealready A2 extremely) and probably dense (Figure A2) and probably do not allow large increases in the building stock. Instead, additional parts do(Figure not allow A2) and large probably increases do in not the allow building large stock. increase Instead,s in the additional building parts stock. or Instead, ﬂoors were additional added to parts the or floors were added to the existing ones. existingor floors ones. were added to the existing ones.

Figure 9. Number of new (c2, left) and changed (c3, right) buildings between 2015 and 2017. Figure 9. NumberNumber of of new new ( (c2,c2, left)left) andand changedchanged (c3,(c3, right) right) buildings buildings between between 2015 2015 and and 2017. 2017. ToTo furthermorefurthermore demonstratedemonstrate thethe valuevalue ofof thethe presentedpresented dataset,dataset, wewe comparedcompared thethe numbernumber ofof To furthermore demonstrate the value of the presented dataset, we compared the number of demolitionsdemolitions displayeddisplayed byby thethe hexagonalhexagonal aggregationaggregation toto thethe demolitionsdemolitions automaticallyautomatically identifiedidentified byby demolitions displayed by the hexagonal aggregation to the demolitions automatically identified by WarthWarth et et al. al. (2019), (2019), whowho derived derived digital digital surface surface models models of the the citycity from from 2015 2015 and and 2017 2017 toto detectdetect Warth et al. (2019), who derived digital surface models of the city from 2015 and 2017 to detect demolitionsdemolitions basedbased on the threshold of the height height difference difference [4] [4].. However, However, they they used used hexagons hexagons with with a demolitions based on the threshold of the height difference [4]. However, they used hexagons with a adiameter diameter of of 500 500 m m,, so so both both datasets datasets were were normalized by calculating the demolitions demolitions per per square square diameter of 500 m, so both datasets were normalized by calculating the demolitions per square kilometer.kilometer. AsAs shownshown inin FigureFigure 1010,, thethe presentedpresented datasetdataset notnot onlyonly ooffersffers finerfiner spatialspatial details,details, whichwhich kilometer. As shown in Figure 10, the presented dataset not only offers finer spatial details, which makemake ititmore more suitable suitable to to describe describe urban urban changes, changes, but but is is also also less less prone prone to to systematic systematic overestimation, overestimation, as make it more suitable to describe urban changes, but is also less prone to systematic overestimation, itas was it was reported reported by Warthby Warth et al. et (2019)al. (2019) [4]. [4]. as it was reported by Warth et al. (2019) [4].

FigureFigure 10.10. DemolitionsDemolitions perper squaresquare kilometerkilometer inin thethe presentedpresented datasetdataset (left)(left) comparedcompared toto ﬁndingsfindings ofof Figure 10. Demolitions per square kilometer in the presented dataset (left) compared to findings of WarthWarth et et al. al. (2019).(2019). Warth et al. (2019). 3.5. Statistical Description

Lastly, we highlight the potential for statistical analyses of the prepared data, which can be conducted based on the attribute tables of the shapeﬁles. Figure 11 aggregates the changes at the Data 2020, 5, x FOR PEER REVIEW 14 of 18

3.5.Data Statistical 2020, 5, x FOR Description PEER REVIEW 14 of 18 3.5. Lastly,Statistical we Description highlight the potential for statistical analyses of the prepared data, which can be conducted based on the attribute tables of the shapefiles. Figure 11 aggregates the changes at the Data 2020Lastly,, 5, 42 we highlight the potential for statistical analyses of the prepared data, which can14 of 18be building level for seven districts covered by the satellite images (please note that Điện Bàn district conducted based on the attribute tables of the shapefiles. Figure 11 aggregates the changes at the was excluded because of very low coverage and small values). It highlights the enormous share of building level for seven districts covered by the satellite images (please note that Điện Bàn district changesbuilding in level the fortotal seven building districts stock covered of the bycity. the satellite images (please note that Điện Bàn district was was excluded because of very low coverage and small values). It highlights the enormous share of excluded because of very low coverage and small values). It highlights the enormous share of changes changes in the total building stock of the city. in the total building stock of the city.

40,000

40,000 30,000

30,000 20,000

20,000 10,000

10,000 0 Cẩm Lệ Hải Châu Hòa Vang Liên Chiểu Ngũ Hành Sơn Sơn Trà Thanh Khê 0 Cẩm Lệ Hải Châu Hòa unchangedVang Liên ChiểuchangedNgũ Hành Sơn Sơn Trà Thanh Khê

unchanged changed Figure 11. Changed and unchanged buildings for the districts in Da Nang. Figure 11. Changed and unchanged buildings for the districts in Da Nang. Figure 11. Changed and unchanged buildings for the districts in Da Nang. Figure 12 shows the proportion of new and demolished buildings based on the hexagon dataset, Figure 12 shows the proportion of new and demolished buildings based on the hexagon dataset, in relation to their average distance to the historic city center of Da Nang. Interestingly, both types of in relationFigure to 12 their shows average the proportion distance to of the new historic and demolished city center ofbuildings Da Nang. based Interestingly, on the hexagon both typesdataset of, changes show a similar pattern, even if they are not displayed in absolute values. Accordingly, 50% changesin relation show to their a similar average pattern, distance even toif the they historic are not city displayed center of inDa absolute Nang. Interestingly, values. Accordingly, both types 50% of of all new buildings and 43% of all demolitions between 2015 and 2017 occurred within a distance of ofchanges all new show buildings a similar and pattern, 43% of even all demolitions if they are betweennot displayed 2015 andin absolute 2017 occurred values. withinAccordingly, a distance 50% three to six kilometers from the historic city center. Within a radius of two kilometers, 12.9% of new ofof threeall new to buildings six kilometers and 43% from of the all historicdemolitions city center.between Within 2015 and a radius 2017 occurred of two kilometers, within a distance 12.9% ofof buildings were built, but 21.2% of demolitions occurred. This again underlines that the central newthree buildings to six kilometers were built, from but the 21.2% historic of demolitions city center. occurred.Within a radius This again of two underlines kilometers, that 12.9% the centralof new residential areas in Da Nang are already very dense and do not allow much more growth. Instead, residentialbuildings areaswere in built, Da Nang but 21.2% are already of demolitions very dense occurred. and do not This allow again much underlines more growth. that Instead, the central the the amount of changed buildings (change type 3, as defined in Table 1) is visibly higher in the more amountresidential of changed areas in buildingsDa Nang(change are already type very 3, as deﬁneddense and in Tabledo not1) isallow visibly much higher more in thegrowth. more Instead, central central areas. Of all changed buildings in Da Nang, 77% were located within the first six kilometers. areas.the amount Of all changedof changed buildings buildings in Da(change Nang, type 77% 3 were, as defined located in within Table the1) is ﬁrst visibly six kilometers. higher in the more central areas. Of all changed buildings in Da Nang, 77% were located within the first six kilometers.

20 %

20 % 15 %

15 % 10 %

10 % 5 %

5 % 0 % 1 km 2 km 3 km 4 km 5 km 6 km 7 km 8 km 9 km 10 km 11 km 12 km 13 km 0 % 1 km 2 km 3 kmnew buildings4 km 5 kmdemolished6 km buildings7 km 8 km changed9 km building10 km 11 km 12 km 13 km

Figure 12. Share of new andnew demolishedbuildings buildingsdemolished with buildings respect to thechanged distance building to the historic city center. Figure 12. Share of new and demolished buildings with respect to the distance to the historic city center. 3.6. Discussions on Data Quality Figure 12. Share of new and demolished buildings with respect to the distance to the historic city center. This section summarizes the most important aspects concerning accuracy. It is obvious that the quality of the presented dataset depends upon two major error sources: the first is the accuracy of all Data 2020, 5, 42 15 of 18 input products, which determines how well the changes between 2015 and 2017 can be identified and classified. If buildings were not identified or classified for 2015, it is likely that an error is propagated into the change detection approach. As reported by Warth et al. [4], the built-up areas of 2015 have been assessed at an accuracy of 88.4%. Accordingly, the actual size of an individual building triggers potential errors. This affects small buildings in the first instance, because they are more sensitive to misclassifications of small image objects. We did not see a propagation of this error into later products, as the quality of building areas was assessed by Vetter-Gindele et al. [3] and reported with 87.2% accuracy. As the subsequent classification of building types was also based on the buildings’ sizes, assessed from the satellite image, incorrect estimates could result in the assignment of a wrong type. This is partly disproved by the confusion matrix shown in Table A1, which shows that most of the misclassifications occurred between buildings of a similar size (e.g., single basic and single local, bungalows and villas, or also multifamily local and multifamily modern). Accordingly, we see a more plausible reason for the incorrect classifications, due to atypical buildings which do not fit into the typology presented in Table1, rather than due to small errors in the size of their footprint. One shortcoming of the of the presented accuracy assessment is the statistical dominance of single family local-type buildings. On the one hand, this type makes up 90% of all buildings in Da Nang and has producer and user accuracy measurements above 95%. On the other hand, the small representation of halls, outbuildings and special buildings allows few conclusions to be drawn about the actual accuracy of these building types, and their low producer accuracies (54.5%, 70.6% and 75.0%), indicate that they are rather underestimated in the dataset in favor of more accurate and complete residential buildings and an overall classification accuracy of 90.9%. However, the chosen approach of visual identification of changes between 2015 and 2017 allowed the interpreters to correct errors in the building type classification of 2015, thereby minimizing misclassifications above the statistically reported accuracy. This leads to the second potential error source of this dataset, which is the subjective interpretation of building types for the image of 2017 and the resulting changes. We tried to minimize the impact of the single interpreter’s perception by distributing this task over a broad range of individuals and by cross-checking random buildings to identify systematic errors. Despite this, interpretation errors can still arise from spectral differences in the two images (Figure3), or in cases where actual changes are overlooked. Accordingly, these errors of omission and commission cannot entirely be avoided. Based on the statistical tests and the considerations above, their impact on the aggregated numbers at an administrative level and the hexagonal grids is considered very small.

Supplementary Materials: The following are available online at: https://doi.org/10.5281/zenodo.3757710. Author Contributions: Conceptualization, A.B., G.W., F.B., T.T.G.B. and V.H.; methodology, A.B., G.W. and F.B.; data generation, A.B., H.T. and F.B.; validation and data curation, A.B.; writing—original draft preparation, A.B.; writing—review and editing, A.B., G.W., F.B., F.B., T.T.G.B., H.T. and V.H.; visualization, A.B.; supervision and funding acquisition, F.B. and V.H.; All authors have read and agreed to the published version of the manuscript. Funding: This research was supported by the German Ministry of Education and Research (Bundesministerium für Bildung und Forschung, BMBF) with the research project Rapid Planning, grant numbers 01LG1301J and 01LG1301K. Acknowledgments: We acknowledge the support of Alice Rühl, Dimitrij Chudinzow, and Sheetal Marathe, who assisted in the collection of the field reference data, and Bianca Deutsch, Carsten Gawlas, Daniel Gruschwitz, Marco Hammer, Sina Pfleiderer, and Sven Endreß, who assisted in the identification of changes. Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data, in the writing of the manuscript, or in the decision to publish the results. Data 2020, 5, x FOR PEER REVIEW 16 of 18

Appendix A. Accuracy Assessment for the Building Dataset of 2015 Data 2020, 5, 42 16 of 18 Table A1. Confusion matrix of building types (BT) between reference data and the classification.

Appendix A. Accuracy Assessment for theReference Building Data Dataset of 2015 BT1 BT2 BT3 BT4 BT5 BT6 BT7 BT8 BT9 sum UA Table A1. Confusion matrix of building types (BT) between reference data and the classiﬁcation. BT1 57 10 0 0 0 0 0 0 3 70 81.4%

BT2 3 519 3 3 Reference0 Data 0 2 0 1 531 97.7%

BT3 2 BT16 BT221 BT3 BT45 BT50 BT60 BT70 BT80 BT90 Sum34 UA 61.8% BT4 BT10 570 104 028 0 01 00 00 00 30 7033 81.4% 84.8% BT5 BT20 30 519 32 348 0 06 20 03 10 53159 97.7% 81.4% BT6 BT30 20 60 210 5 04 027 00 02 00 3433 61.8% 81.8% BT4 0 0 4 28 1 0 0 0 0 33 84.8%

Classification BT7 0 0 0 0 0 0 6 0 0 6 100.0% BT5 0 0 2 48 6 0 3 0 59 81.4% BT8 BT60 00 00 01 0 40 270 02 212 00 3315 81.8% 80.0%

BT9Classiﬁcation BT73 05 01 00 0 0 00 61 00 012 6 22100.0% 54.5% sum BT865 0540 029 039 1 53 0 033 211 1217 016 15803 80.0% PA BT987.7% 396.1% 572.4% 1 71.8% 0 90.6% 0 81.8% 0 54.5% 1 070.6% 1275.0% 22 54.5%90.9% BT sum= building 65 types 540(1 = single 29 basic; 2 39 = single 53 local; 3 33= single 11bungalow; 17 4 = single 16 villa; 803 5 = multi PA 87.7% 96.1% 72.4% 71.8% 90.6% 81.8% 54.5% 70.6% 75.0% 90.9% local; 6 = multi modern; 7 = special; 8 = hall; 9 = outbuilding), PA = producer's accuracy, UA = user's BTaccuracy= building. types (1 = single basic; 2 = single local; 3 = single bungalow; 4 = single villa; 5 = multi local; 6 = multi modern; 7 = special; 8 = hall; 9 = outbuilding), PA = producer’s accuracy, UA = user’s accuracy. Appendix B. Additional Maps Appendix B. Additional Maps

Figure A1. Share of each ward covered by the area of interest (AOI, extent of the satellite images) of Figure A1. Share of each ward covered by the area of interest (AOI, extent of the satellite images) of this study. this study.

DataData2020 2020,,5 5,, 42 x FOR PEER REVIEW 1717 of 1818 Data 2020, 5, x FOR PEER REVIEW 17 of 18

FigureFigure A2.A2. BuildingBuilding densitydensity inin 20172017 aggregatedaggregated byby hexagons.hexagons. Figure A2. Building density in 2017 aggregated by hexagons.

FigureFigure A3.A3. AbsoluteAbsolute numbernumber ofof newnew buildingsbuildings comparedcompared toto areasareas expectedexpected toto bebe belowbelow thethe tidelinetideline inin Figure20502050 byby A3 KulpKulp. Absolute && StraussStrau numberss (2019)(2019) of [[19] 19new].. buildings compared to areas expected to be below the tideline in 2050 by Kulp & Strauss (2019) [19]. ReferencesReferences References 1.1. Zhang,Zhang, B.-G.B.-G. ApplicationApplication ofof remoteremote sensingsensingtechnology technologyto topopulation population estimation. estimation.Chin. Chin. Geogr.Geogr. Sci.Sci. 20032003,, 13,, 1. Zhang,267–271.267–271. B. [-CrossRefG. Application] of remote sensing technology to population estimation. Chin. Geogr. Sci. 2003, 13, 2.2. 267Xie,Xie,– Y.;271.Y.; Weng,Weng, A.; A.;Weng, Weng, Q. Q. Population Populationestimation estimationof of urbanurban residentialresidentialcommunities communities usingusing remotelyremotely sensedsensed

2. Xie,morphologicmorphologic Y.; Weng, data. data.A.; Weng, IEEEIEEE Geosci.Q.Geosci. Population Remote estimation Sens. Lett. 2015 2015of urban,, 12,, 1111 1111–1115.residential–1115. communities using remotely sensed 3.3. morphologicVetter-Gindele,Vetter-Gindele, data. J.;J.; Braun,IEEEBraun, Geosci. A.;A.; Warth,Warth, Remote G.;G.; Sens. Bui,Bui, Lett. T.T.Q.;T.T.Q.; 2015Bachofer, Bachofer,, 12, 1111 F.; F.;–1115. Eltrop, Eltrop, L. L. Assessment Assessmentof ofHousehold Household SolidSolid 3. VetterWasteWaste- Generation Gindele,Generation J.; andBraun, and Composition CompositionA.; Warth, by G.; Buildingby Bui, Building T.T.Q.; Type Bachofer, inType Da Nang,in F.;Da Eltrop, Vietnam. Nang, L. Vietnam. ResourcesAssessment Resources2019 of, 8Household, 171. 2019 [CrossRef, 8 Solid, 171,] 4. WasteWarth,doi:10.3390/resources8040171. G.;Generation Braun, A.; and Bödinger, Composition C.; Hochschild, by Building V.; Bachofer, Type in F. DSM-basedDa Nang, Vietnam. identiﬁcation Resources of changes 2019 in, 8 highly, 171, 4. doi:10.3390/resources8040171.dynamicWarth, G.; urban Braun, agglomerations. A.; Bödinger, Eur. C.; Hochschild, J. Remote Sens. V.;2019 Bachofer,, 52, 322–334. F. DSM [CrossRef-based identification] of changes in 4.5. Warth,Unitedhighly Nations.G.; dynamicBraun, World A.; Bödinger,urban Population agglomerations.C.; Prospects: Hochschild, Revision V.;Eur. Bachofer, of 2019. J. F. Available Remote DSM- based online: Sens. identification https:2019//population.un.org, 52of, changes322–334, in/ highlywppdoi:10.1080/22797254.2019.1604083./ (accessed dynamic on 23 Februaryurban 2020).agglomerations. Eur. J. Remote Sens. 2019, 52, 322–334, 5. doi:10.1080/22797254.2019.1604083.United Nations. World Population Prospects: Revision of 2019. Available online: 5. Unitedhttps://population.un.org/wpp/ Nations. World Population(accessed on 23Prospects: February 2020).Revision of 2019. Available online: https://population.un.org/wpp/ (accessed on 23 February 2020).

Data 2020, 5, 42 18 of 18

6. Ostojic, D.R.; Bose, R.K.; Krambeck, H.; Lim, J.; Zhang, Y. Energizing Green Cities in Southeast Asia; The World Bank: Washington, DC, USA, 2013; ISBN 978-0-8213-9837-1. 7. Burnett, M.G. Hexagon (KH-9). Mapping Camera Program and Evolution; United States Secretary of the Air Force: Washington, DC, USA, 2012. 8. Moon, K.; Downes, N.; Rujner, H.; Storch, H. Adaptation of the urban structure type approach for vulnerability assessment of climate change risks in Ho Chi Minh City. In Congress Report, Low Carbon Cities; International Planning Congress: Porto, Portugal, 2009. 9. Downes, N.K.; Storch, H.; Schmidt, M.; Nguyen, T.C.V.; Tran, T.N. Understanding Ho Chi Minh City’s urban structures for urban land-use monitoring and risk-adapted land-use planning. In Sustainable Ho Chi Minh City: Climate Policies for Emerging Mega Cities; Katzschner, A., Waibel, M., Schwede, D., Katzschner, L., Schmidt, M., Storch, H., Eds.; Springer: Heidelberg, Germany, 2016; pp. 89–116. ISBN 978-3-319-04615-0. 10. Bachofer, F.; Braun, A.; Adamietz, F.; Murray, S.; d’Angelo, P.; Kyazze, E.; Mumuhire, A.P.; Bower, J. Building Stock and Building Typology of Kigali, Rwanda. Data 2019, 4, 105. [CrossRef] 11. GADM. Database of Global Administrative Areas: Version 3.6. Available online: https://gadm.org (accessed on 15 January 2020). 12. Birch, C.P.D.; Oom, S.P.; Beecham, J.A. Rectangular and hexagonal grids used for observation, experiment and simulation in ecology. Ecol. Model. 2007, 206, 347–359. [CrossRef] 13. Canny, J. A computational approach to edge detection. IEEE Trans. Pattern Anal. Mach. Intell. 1986, 679–698. [CrossRef] 14. Panagiotakis, E.; Chrysoulakis, N.; Charalampopoulou, V.; Poursanidis, D. Validation of Pleiades Tri-Stereo DSM in urban areas. ISPRS Int. J. Geo Inf. 2018, 7, 118. [CrossRef] 15. General Statistics Oﬃce of Vietnam. The 2009 Vietnam Population and Housing Census: Major Findings. Available online: https://www.gso.gov.vn/default_en.aspx?tabid=515&idmid=5&ItemID=9813 (accessed on 15 January 2020). 16. Hằng, N.H. Da Nang city development. In Knowledge Creation in Community Development; Nishihara, A.H., Matsunaga, M., Nonaka, I., Yokomichi, K., Eds.; Springer: Heidelberg, Germany, 2018; pp. 85–106. 17. Bachofer, F.; Rau, H. Change monitoring of a heterogeneous urban landscape using RapidEye data—DaNang, Vietnam. In Proceedings of the GIS 2015 Conference, Hue, Vietnam, 8–11 December 2016; pp. 702–707. 18. Linh, N.H.K.; Chuong, H.V. Assessing the impact of urbanization on urban climate by remote sensing perspective: A case study in DaNang city, Vietnam. Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. 2015, XL-7/W3, 207–212. [CrossRef] 19. Kulp, S.A.; Strauss, B.H. New elevation data triple estimates of global vulnerability to sea-level rise and coastal ﬂooding. Nat. Commun. 2019, 10, 4844. [CrossRef][PubMed]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).