DETERMINING EMPLOYMENT OPPORTUNITY AREAS IN AFRICAN CITIES A SCALABLE RAPID APPRAISAL METHODOLOGY - CASE STUDY OF CITY

ISSUE INDICATORS KAMPALA CITY EMPLOYMENT PROFILE • Access to people, goods, services, and information is the basis of economic The indicator datasets were derived from three main sources: N development in cities. The better and more eff icient this access, the greater the • Crowd Sourced - Open Streets Map/Google Maps/FSP Mapping • Kampala is the largest city in economic benefi ts through economies of scale, agglomeration eff ects, and net- • Satellite Imagery Derivatives - Open Streets Map/ESRI/World Bank • Greater Kampala Metropolitan Area is the main center of economic activity MAKARERE METROPLEX UNIVERSITY COLLEGE SHOPPING working advantages. Cities facilitate accessibility through their urban forms and • Local City Authorities: Land Use Policy Mapping etc. MALL within the country. NATETE KISEMENTI KIGOWA OLD KAMPALA transport systems impacts directly on other measures of human development ALA KYAMBOGOKYAMBO UNIVERSITY CITYCITC Y CENTERCEN MAKARERE UNIVERSU • Employment is heavily concentrated in the urban core. LUMLUMBIRIU BIRI UNIVERSITY and well-being. These were translated into the following list of indicators: NDEEBA BUSINESS SCHOOL INDUSTRIAL AREA MBUYAMBUM YA HOSPITAL H • CBD Employment densities - 5000 and 10,000 workers/km2 KAMPALA INTERNATIONAL • Adequate access to spatially dispersed resources (consumers, jobs, suppliers) UNIVERSITY • 55% of all fi rms and 62% of all jobs within 5km of the CBD. are vital for fi rms and households to thrive or even just to survive. Most urban transport projects aim at enhancing citizens’ ‘mobility’ (buff er and travel time) • Most jobs near the CBD are in the public sector (e.g., government) and in the instead of enhancing a citizen’s access to opportunities. tradable sector (e.g., manufacturing, business services). GABA • This has lead cities to develop transportation infrastructure that have mostly • 57% of manufacturing jobs and 80% of business services jobs within 5km of bred more congestion and the perils that come with it (e.g. auto-dependency; the CBD. longer travel times for the poor vis-à-vis higher income counter parts; pollution; climate change). • Dysfunctional transport and land markets have led to cities failing to capture the productivity and agglomeration benefi ts of integrated labor markets that Projected spatial distribution of employment by Traff ic Analysis Zones (TAZs) are common in many cities. translated into 500m cell grid for comparison with alternative indicator data- sets. • At the core, most cities are failing in connecting people eff iciently with job op- portunities and other places of interest. Many people, but especially the poor, suff er a lack of access and high costs for transport to jobs and other opportu- nities, a concern exacerbated by the particular economic geography in Africa. N

KAWEMPE Accessibility off ers a powerful lens to assess how a mobility system

MAKARERE METROPLEX is serving an urban area. It can reframe the eff iciency of our trans- UNIVERSITY COLLEGE SHOPPING MALL NATETE KISEMENTI KIGOWA OLD KAMPALA port systems in terms of their ability to connect people with op- ALA KYAMBOGOKYAMBO UNIVERSITY CITYCITC Y CENTERCEN MAKARERE UNIU VERS LUMBIRILUMU BIRI UNIVERSITY portunities rather than mobility. In parallel, the use of counterfac- NDEEBA BUSINESS SCHOOL

INDUSTRIAL AREA MBUYAM HOSPITAL H KAMPALA tual scenarios can help identify land use planning options that can INTERNATIONAL UNIVERSITY maximize the connectivity impact of transport investments. Utiliz- ing new tools to study accessibility therefore off ers a lens by which LUZIRA one can better understand and plan transport and land use policies. GABA

80 % Correlation ( 72% outside CBD) Combined Indicators: Open Street Maps Amenities CHALLENGE Banking locations count within cell & adjacent cells (no mobile money agents) Planned Employment zoned area The main challenge that remains to globalize this work is getting spatialized eco- (cells with greater than 10 Ha of employment zoned land use) nomic opportunity data. • In most cases, such data is outdated, non-existent, or only capture areas of RESULTS Combined Indicators (Banking and Planned Employment Zones) showing an formal economic activity, leaving out signifi cant numbers of informal employ- The results of the Kampala City test case were quite surprising showing up to 80 % correlation to actual employment distribution data for the 80% Correlation to TAZ employment distribution. ment opportunities, which in most African Cities, would be a large portion of a city. The key factor was the mapping of amenities and banks. The ratio between the city population and the number of amenities was 0.56. city’s economy. With such high correlation, the Bank has asked to run the test on two other Cities - Dakar & Nairobi. Pending the results from those, seven • To counter this lack of information, alternative datasets that refl ect employ- other cities will done, where verifi cation data is not available. The limitations of this methodology are primarily the availability of adequate ment activity can be used to determine employment opportunity areas(EO- crowd-sourced data. For cities with low internet penetration, it would be diff icult to achieve such high accuracy. N

As). KAWEMPE • Economic activity physically and temporally, impacts the locations where it ŽƌƌĞůĂƚŝŽŶ^ĐŽƌĞƐ takes place, therefore there would be other indicators that help estimate the MAKARERE METROPLEX UNIVERSITY COLLEGE SHOPPING World Bank Collected Business Registry Data Ϭ͘ϱϴ Ϭ͘ϲϳ Ϭ͘ϱϲ Ϭ͘ϲϬ Ϭ͘ϲϲ Ϭ͘ϲϱ Ϭ͘ϲϴ Ϭ͘ϲϵ Ϭ͘ϲϱ Ϭ͘ϳϬ Ϭ͘ϲϴ Ϭ͘ϲϴ Ϭ͘ϳϭ Ϭ͘ϲϰ Ϭ͘ϲϱ Ϭ͘ϲϳ Ϭ͘ϲϯ Ϭ͘ϲϮ Ϭ͘ϲϱ Ϭ͘ϲϴ Ϭ͘ϲϴ Ϭ͘ϳϬ Ϭ͘ϱϵ Ϭ͘ϳϱ Ϭ͘ϲϮ Ϭ͘ϳϵ Ϭ͘ϯϳ Ϭ͘ϱϰ Ϭ͘Ϯϴ Ϭ͘ϯϵ Ϭ͘ϭϵ Ϭ͘Ϭϲ MALL locations of such activity. NATETE KISEMENTI KIGOWA OLD KAMPALA KCCA 2018 Projected Employment Data by ALA KYAMBOGOKYAMBO Ϭ͘ϴϬ Ϭ͘ϴϬ Ϭ͘ϳϵ Ϭ͘ϳϵ Ϭ͘ϳϵ Ϭ͘ϳϵ Ϭ͘ϳϵ Ϭ͘ϳϵ Ϭ͘ϳϴ Ϭ͘ϳϴ Ϭ͘ϳϴ Ϭ͘ϳϴ Ϭ͘ϳϴ Ϭ͘ϳϳ Ϭ͘ϳϳ Ϭ͘ϳϱ Ϭ͘ϳϭ Ϭ͘ϳϭ Ϭ͘ϳϬ Ϭ͘ϲϵ Ϭ͘ϲϴ Ϭ͘ϲϳ Ϭ͘ϲϱ Ϭ͘ϲϯ Ϭ͘ϱϵ Ϭ͘ϱϴ Ϭ͘ϱϳ Ϭ͘ϱϱ Ϭ͘ϱϭ Ϭ͘ϰϳ Ϭ͘ϯϵ ͲϬ͘Ϭϯ UNIVERSITY Traffc Analysis Zones CITYCITC Y CENTERCEN MAKARERE UNIU VERS • With the proliferation of mobile phone technology, social media, and crowd LUMBIRILUMU BIRI UNIVERSITY NDEEBA BUSINESS SCHOOL

INDUSTRIAL AREA MBUYAM HOSPITAL H sourced spatial mapping platforms, the ability to utilize these data sources in- Open Streets Map Amenities Data z zzzzzzzzz z z zz zz KAMPALA INTERNATIONAL UNIVERSITY dividually, or in correlation with each other, a robust methodology to estimate Banking locations count within cell & the location and size of EOAs using these data sources would be very useful adjacent cells zzzzzzzzzzzzzzzz LUZIRA (no mobile money agents)

in rapidly appraising the locations of EOAs in emerging cities. Banking locations count GABA z (no mobile money agents)

Finance Access Location count within cell & 80 % Correlation ( 72% outside City Center) zzzzz The project involved testing the assumption that alternative data- adjacent cell Heights of Cells in comparison with

sets refl ecting employment activity could accurate enough to deter- Finance Access Location count z KCCA 2018 TAZ Employment Projections mine the spatial distribution of employment - identifying EOAs. For Color of Cells the test, the City of Kampala was selected as the World Bank has ac- Average Building Footprint Size z Average Building Footprint Size cess to spatialized economic opportunity data, to verify the accuracy z zzzz zz zzz zz z A visual comparison the combined indicators (height) to the TAZ employment of the alternative datasets. ( twice the city mean) (color) showing a fairly accurate representation of the spatial distribution of Building Cover zz z z Jobs. THEORETICAL APPROACH Building Density • Gather readily available data sources that would either be physical impacts Slum Location z CLIENT: of economic activity (density, building sizes, better connectivity, organic con-

ǀĂŝůĂďůĞ/ŶĚŝĐĂƚŽƌĂƚĂƐĞƚƐ Slum Location z z TRANSPORT & ICT GLOBAL PRACTICE centrations of supportive activity etc.), or be a demographic representation (greater than 10 Ha in size) of employment (businesses/amenities location concentrations). Planned Employment Land use coverage z z • Compare gathered datasets to actual employment distribution data to deter- Planned Employment Land use coverage zz zz mine which datasets or combination thereof would be most accurate. (over 10 Ha of employment zoned area) • Identify key benchmarks for gathered datasets that can be scalable to other Intersection Proximity z zzz z z z z CONSULTANT: cities where economic data is unreliable. (cells above mean only) Proximity to Bus/Taxi Rank zzzzz z (cells above mean only) BHARAT SINGH, MDP 2018

World Bank Collected Business Registry Data Ϭ͘ϰϯ Ϭ͘ϱϵ Ϭ͘ϰϯ Ϭ͘ϰϴ Ϭ͘ϲϮ Ϭ͘ϱϲ Ϭ͘ϲϬ Ϭ͘ϱϴ Ϭ͘ϱϳ Ϭ͘ϲϲ Ϭ͘ϲϬ Ϭ͘ϱϵ Ϭ͘ϲϲ Ϭ͘ϱϳ Ϭ͘ϱϱ Ϭ͘ϲϭ Ϭ͘ϰϯ Ϭ͘ϱϭ Ϭ͘ϲϬ Ϭ͘ϱϵ Ϭ͘ϱϵ Ϭ͘ϱϰ Ϭ͘ϰϬ Ϭ͘ϳϭ Ϭ͘ϱϲ Ϭ͘ϳϱ Ϭ͘ϰϳ Ϭ͘ϰϱ Ϭ͘Ϯϴ Ϭ͘ϱϭ ͲϬ͘Ϭϭ Ϭ͘Ϯϱ https://www.linkedin.com/in/bartsingh Background Image: https://commons.wikimedia.org/wiki/File:Overview_of_Minibuses_in_Taxi_Park_-_Downtown_Kampala_-_Uganda.jpg KCCA 2018 Projected Employment Data by Ϭ͘ϲϱ Ϭ͘ϳϮ Ϭ͘ϲϴ Ϭ͘ϱϵ Ϭ͘ϳϮ Ϭ͘ϳϰ Ϭ͘ϲϴ Ϭ͘ϲϱ Ϭ͘ϳϯ Ϭ͘ϲϲ Ϭ͘ϲϵ Ϭ͘ϲϵ Ϭ͘ϲϯ Ϭ͘ϲϱ Ϭ͘ϲϭ Ϭ͘ϲϰ Ϭ͘ϲϭ Ϭ͘ϳϭ Ϭ͘ϲϲ Ϭ͘ϲϬ Ϭ͘ϲϰ Ϭ͘ϱϱ Ϭ͘ϰϰ Ϭ͘ϱϮ Ϭ͘ϱϭ Ϭ͘ϰϲ Ϭ͘ϲϭ Ϭ͘ϱϱ Ϭ͘ϱϱ Ϭ͘ϲϭ Ϭ͘Ϯϵ Ϭ͘Ϭϲ Traffc Analysis Zones ŽƌƌĞůĂƚŝŽŶ^ĐŽƌĞƐĞdžĐůƵĚŝŶŐĞŶƚƌĂůƵƐŝŶĞƐƐŝƐƚƌŝĐƚ