US 201400.40371 A1 (19) United States (12) Patent Application Publication (10) Pub. No.: US 2014/0040371 A1 Gurevich et al. (43) Pub. Date: Feb. 6, 2014
(54) SYSTEMIS AND METHODS FOR filed on Jun. 14, 2010, provisional application No. IDENTIFYING GEOGRAPHC LOCATIONS 61/354,556, filed on Jun. 14, 2010, provisional appli OF SOCIAL MEDIA CONTENT COLLECTED cation No. 61/354,559, filed on Jun. 14, 2010, provi OVER SOCIAL NETWORKS sional application No. 61/617,524, filed on Mar. 29, 2012, provisional application No. 61/618,474, filed on (71) Applicants: Olga Gurevich, San Francisco, CA Mar. 30, 2012. (US); Rishab Aiyer Ghosh, San Francisco, CA (US) Publication Classification (72) Inventors: Olga Gurevich, San Francisco, CA (51) Int. Cl. (US); Rishab Aiyer Ghosh, San H04L 29/2 (2006.01) Francisco, CA (US) (52) U.S. Cl. CPC ...... H04L 61/609 (2013.01) (73) Assignee: TOPSY LABS, INC., San Francisco, USPC ...... 709/204 CA (US) (57) ABSTRACT (21) Appl. No.: 13/853,687 A new approach is proposed that contemplates systems and methods to identify geographic locations of all Social media (22) Filed: Mar. 29, 2013 content items retrieved form a social network in real time, wherein the geographic locations are physical locations from Related U.S. Application Data which the Social media content items are originated or (63) Continuation-in-part of application No. 13/158,992, authored. If the latitude/longitude (geographic) coordinates filed on Jun. 13, 2011, which is a continuation-in-part of the content item are available, the geographic location of of application No. 12/895,593, filed on Sep. 30, 2010, the Social media content item can be identified using Such now Pat. No. 7,991,725, which is a continuation-in geographic coordinates. For content items which geographic part of application No. 12/628,801, filed on Dec. 1, coordinates are not available, historical archive of content 2009, now Pat. No. 8,244,664, which is a continuation items with high-confidence of geographic locations are uti in-part of application No. 12/628,791, filed on Dec. 1, lized to train a location classifier to predict geographic loca 2009. tions of Such content items with high accuracy. Finally, the identified locations of the content items are confirmed to be (60) Provisional application No. 61/354.551, filed on Jun. accurate and are presented to a user together with the content 14, 2010, provisional application No. 61/354,584, items.
100
Author One Citation i. 1 1 tie (influence Score 10)
related to
:
; AuthOrt/O COmpOSes (influencescore 5) Citation 21 - larget Two
Atitior Three H Citation 3.2 - 4 ties (influence Score 4)
3. ^. ------. ---. a 102 i04 106 Patent Application Publication Feb. 6, 2014 Sheet 1 of 19 US 2014/0040371 A1
sau??.
***************** S9SOdulo0
`-001 Patent Application Publication Feb. 6, 2014 Sheet 2 of 19 US 2014/0040371 A1
Patent Application Publication Feb. 6, 2014 Sheet 3 of 19 US 2014/0040371 A1
SOCIAL INTELLIGENCESYSTEM20 quy Dashboard Admin Settings Sign Out Syria topic
(clamascus-toms:re-rrxM------srasia: Twitters last 24 hours
F. 3
Save atopic Maitle (e) ReplaceStratopic an existing topic mawawaaxxxxx------O Create a new Topic
KeywordsEdamascus homs -a-ra-rarasyria ------Syria Filters ?last 24 hours Twitter
Patent Application Publication Feb. 6, 2014 Sheet 4 of 19 US 2014/0040371 A1
rearm
save as... Tornado v. Filter by name Twitter (Last 24 Hours) Iran Middle East Putin Syria Tornado Obama
(C) Locations O languageS
anaaaaaaaaaaaaaaaaa. Q Sentiment
Source reneuraenaeanananaaaaaaaaaaa. 3-7 influence
FG. 6 Patent Application Publication Feb. 6, 2014 Sheet 5 of 19 US 2014/0040371 A1
DateS
LaSt hOur LaSt 24 hours last 7 dayS last 30 dayS Last 90 dayS
Last 180 days Specific date range.
seracre-enaaaaaaaaaaa. Start Date
End Date
LOCationS
All Locations Search for locationS.
Search for a location ReStrict to lat/long Patent Application Publication Feb. 6, 2014 Sheet 6 of 19 US 2014/0040371 A1
languageS
All Languages
Japanese Korean RuSSian
serremann:
Q Sentiment Positive
Neutral Negative
3-7 influence -- O influential Only Patent Application Publication Feb. 6, 2014 Sheet 7 of 19 US 2014/0040371 A1
F. 2 Patent Application Publication Feb. 6, 2014 Sheet 8 of 19 US 2014/0040371 A1
is: ------a st ticks a hisAll iyila Related terms: Fialkee andley bogi Bogi callelo howard trade dight hoard trade inhalkee Bucks digit Roward 2 FIG. 13
On the road in Milwaukeetonight lets get this One #Knicks! Mar 9, 2012 iO40AM Qstalley Stalley Good morning LA #LinSanity httpitcOlglF49dy
Galenrose JALENR0SE My2 cents, if Howard got a "star playerto COmeplay whim in Orlando he could contend for a title yearly. He is that good #NBA Mar 8, 2012748 PM 7.) G tootiebooty Slim
i? BA PARTTWO http:liCO2aLALPFH ye Mar 9, 20123.17 PM FIG. 14 Patent Application Publication Feb. 6, 2014 Sheet 9 of 19 US 2014/0040371 A1
Top Links () TITLE NBA-Hoya las 9:30pm, (18-21) vs. Milwaukee (15-24), Véalo. Mar 9, 20123:20 PM Milwaukee Bucks vs New York Knicks Live StreamOnline Mar 9, 2012320PM Cuban Calls Own Statement totally SophOmOric,...#NBA #Dallas. MarKen Berger-CBSSports.com9, 2012320PM Trade Deadline Update wnwr------'3-Y-vs.--v BostonMar 9, 20122:37PMmight trade one of "Big Three" butprice is steep------rrrrrrrrrrrrrr------#PBT. Mar 9, 2012242PM WatchLive Cleveland Cavaliers-Oklahoma City Thunder Online. Mar 9, 20123:18PM FIG. 15
Top Mediac
ge
i. i8 Mar 8, 2012.5:26 PM 'i' i... I SC lar 9. 202220 P. it. Mar 9, 2012 1230 PM tick Mar 9, 2012 I-46AA Mars, 2012.5:08AM
FG. f6 Patent Application Publication Feb. 6, 2014 Sheet 10 of 19 US 2014/0040371 A1
Patent Application Publication Feb. 6, 2014 Sheet 11 of 19 US 2014/0040371 A1
•------?áÃ?------~~~~~--~~~~ ~~~~~~~~~~~~~ ----~~~
Patent Application Publication Feb. 6, 2014 Sheet 12 of 19 US 2014/0040371 A1
ZOGENZOGroti N i i i Yt 2am 5am 8am 11am 2pm tornado
Top Posts IIS arrarasarracea------wes
a. al-a-ra-ra- aaaaaaaaaaaa to posts for 7pm-6pm aboutina - Galerose IIE ROSE listian2ns had a star player coreply in noncole oil cond raileyer. He is a good #E4 Coroloskeln Moliseo #NBA loyios is estan Pas Palos Gelas Papas Paadas Quedabar ene coiledo eSGolar, Tular 8740pm
gasketiaido aasters 93-94 base Chicago lasties Blis is It is 874. I Gnosports NBC Sports T 8 8 3. Alagic make is winning streak disapear hiol colonkSmoko inha image this "Thi Mar 8749pi
Bobasketballak Kirtlein . Derick Rose is list like yoy-sick of Dwigh. Howard talk III coli. Clov iPBT#NBA "I a 378m Son ore Patent Application Publication Feb. 6, 2014 Sheet 13 of 19 US 2014/0040371 A1
{
%
44 4 4 4 „…º…~~~~~…,~~~~…... • (S) S. ••••••••••••••••••••••••••••••••••••••••••••••••••••••••¿•~~~~~~~~~~~~~~~~~~*~~~~~~~~~#~ ^^^^^^^^}}}|}} (S
~~~~!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!~~~~~~~•~~~~~%!!!!~~~~~.~~~~.~~~~.~~~~~.~~~~~~~~~~~~ s
e e St. St.
es sae;####;mae ar
3. Patent Application Publication Feb. 6, 2014 Sheet 14 of 19 US 2014/0040371 A1
xxxxxxx xxxxxssr-rrass. --asaaaaaaaaaaaaaaaanxxxxxxx xxxx w &
w S w
Y +++++~~~~)*?--~~~~~~.~~~~~~~~~);
--~~~~~~~~~~~~~~~~~);
Patent Application Publication Feb. 6, 2014 Sheet 15 of 19 US 2014/0040371 A1
------M-MCs:Failer last 24 hours)
S&S&$888&s ------f issisi 888 per 888 & 8888.888 &xe's . - ... ?.
ises ww.rrrrrrrrrm-www.M-WW------v-v-v------xx-& 8 &isis is 888,
RE is sixth radaisode it is is
Saxias & & i8, the Rhisig bass 385;an $8:a::sec3: $88 Six: 8& is isssssss clos Psi S. 2 & S
lifetime of originaling post ------() Mumumwawaaaaaaaaaaaawaaaawawaxx arrarare-as-s------S19$3. Sysissississistis.&sitesis&s over Maukee willout Stoudemies thin the neig, Alek scoredapis, Davis had 7 assists. 10 :88):33:3R8
RI:
Patent Application Publication Feb. 6, 2014 Sheet 17 of 19 US 2014/0040371 A1
alrearsiyara
Patent Application Publication Feb. 6, 2014 Sheet 18 of 19 US 2014/0040371. A
82"{}}}
:? : •• • * --~~~~~~~~~~~~ºº-- A a
arra-a-a- xxxarrrr, Patent Application Publication Feb. 6, 2014 Sheet 19 of 19 US 2014/0040371 A1
Retrieve all SOCial media COntent itemS from a SOCial network in real time, Wherein the retrieved SOCial media COnient itemS do not have IP address identifiers to identify their geographic locations 2902
Identity and mark continuously each of the SOCial media COntent itemS retrieved in real time With its proper geographic location from which Such Content item is authorized Or Originated 2904
Confirm the identified location of a content item by Comparing the identified location to locations of prior Content items posted or authorized by the Same author 2906
SOCial media COntent itemS With their aSSOCiated geographic locationS to a USer 2908
FIG. 29 US 2014/0040371 A1 Feb. 6, 2014
SYSTEMS AND METHODS FOR websites, public conversation repositories, Internet Chat sys IDENTIFYING GEOGRAPHC LOCATIONS tems) generally do not provide IP addresses for their users OF SOCIAL MEDIA CONTENT COLLECTED posts in real-time APIs available to third parties. OVER SOCIAL NETWORKS 0011. The foregoing examples of the related art and limi tations related therewith are intended to be illustrative and not RELATED APPLICATIONS exclusive. Other limitations of the related art will become 0001. This application is a continuation-in-part of current apparent upon a reading of the specification and a study of the copending U.S. application Ser. No. 13/158,992 (Attorney drawings. Docket No.TPY 0001) filed Jun. 13, 2011, which claims the BRIEF DESCRIPTION OF THE DRAWINGS benefit of U.S. Provisional Patent Application No. 61/354, 551, 61/354,584, 61/354,556, and 61/354,559, all filed Jun. 0012 FIG. 1 depicts an example of a citation diagram 14, 20010. U.S. application Ser. No. 13/158,992 is also a comprising a plurality of citations. continuation in part of U.S. Pat. No. 7.991,725 issued Aug. 2, 0013 FIG. 2 depicts an example of a system diagram to 2011 (Attorney Docket No. TPY 0013C1), a continuation in Support interactive presentation and analysis of content over part of U.S. Pat. No. 8,244,664 issued Aug. 14, 2012 (Attor Social networks. ney Docket No. TPY 0017), and a continuation in part of 0014 FIG. 3 depicts an example of a user interface used current copending U.S. application Ser. No. 12/628,791 (At for conducting a query search over a Social network. torney Docket No. TPY 0014) filed Dec. 1, 2009. 0015 FIG. 4 depicts an example of a user interface used 0002 This application claims the benefit U.S. Provisional for saving a search topic over a social network. Patent Application No. 61/617,524, filed Mar. 29, 2012, and 0016 FIG. 5 depicts an example of a dropdown menu for entitled “Social Analysis System,” and is hereby incorporated saved search topics over a Social network. herein by reference. 0017 FIG. 6 depicts an example of a plurality of param 0003. This application claims the benefit of U.S. Provi eters used to refine search results over a Social network. sional Patent Application No. 61/618,474, filed Mar. 30. 0018 FIG. 7 depicts an example of time ranges used to 2012, and entitled “GEO-Tagging Enhancements.” and is refine search results over a social network. hereby incorporated herein by reference. 0019 FIG. 8 depicts an example of locations used to refine search results over a Social network. BACKGROUND 0020 FIG.9 depicts an example of language options used 0004 Social media networks such as Facebook(R), Twit to refine search results over a social network. ter(R), and Google Plus(R) have experienced exponential 0021 FIG. 10 depicts an example of sentiments used to growth in recently years as web-based communication plat refine search results over a social network. forms. Hundreds of millions of people are using various 0022 FIG. 11 depicts an example of user influence used to forms of Social media networks every day to communicate refine search results over a social network. and stay connected with each other. Consequently, the result 0023 FIG. 12 depicts an example of an attention diagram ing activities/contentitems from the users on the Social media used to measure user influence over a social network. networks, such as tweets posted on Twitter R, become phe 0024 FIG. 13 depicts an example of an activity snapshot nomenal and can be collected for various kinds of measure related to keywords mentioned over a period of time on a ments, presentation and analysis. Specifically, these user Social network. activity data can be retrieved from the social data sources of 0025 FIG. 14 depicts an example of a snapshot of top the social networks through their respective publicly avail posts related to keywords mentioned on a social network. able Application Programming Interfaces (APIs), indexed, 0026 FIG. 15 depicts an example of a snapshot of top processed, and stored locally for further analysis. links related to keywords mentioned on a social network. 0005. These stream data from the social networks col 0027 FIG. 16 depicts an example of a snapshot of top lected in real time, along with those collected and stored over media related to keywords mentioned on a social network. time, provide the basis for a variety of measurements, presen 0028 FIG. 17 depicts an example of a snapshot of activi tation and analysis. Some of the metrics for measurements ties associated with keywords overa period of time on a Social and analysis include but are not limited to: network. 0006 Number of mentions Total number of mentions (0029 FIG. 18 depicts an example of share of voice (SOV) for a keyword, term or link: analysis of activities over a period of time on a social network. 0007 Number of mentions by influencers Total num 0030 FIG. 19 depicts an example of Zooming in and out ber of mentions for a keyword, term or link by an influ on activities over a period of time on a social network. ential user; 0031 FIG. 20 depicts an example of viewing of top posts 0008 Number of mentions by significant posts Total with a specific time range on a Social network. number of mentions for a keyword, term or link by 0032 FIG. 21 depicts an example of viewing of keywords/ tweets that have been re-tweeted or contain a link; terms mentioned over a period of time on a Social network. 0009 Velocity. The extent to which a keyword, term 0033 FIG. 22 depicts an example of content items related or link is “taking off in the preceding time windows to discovered terms over a period of time on a social network. (e.g., seven days). 0034 FIG. 23 depicts an example of top trending posts 0010 Unlike traditional web traffic sources, social media over a period of time on a Social network. content items such as citations/Tweets/posts do not have IP 0035 FIG.24 depicts an example of the activities of a post address identifiers to identify their geographic locations. In over its lifetime on a social network. addition, for privacy reasons, social media sources or social 0036 FIG. 25 depicts an example of top trending links networks (such as Twitter R, Facebook.(R), Blogs, review sites, over a period of time on a Social network. US 2014/0040371 A1 Feb. 6, 2014
0037 FIG. 26 depicts an example of top trending media make citations. Each subject 102 or object 106 in diagram 100 over a period of time on a Social network. represents an influential entity, once an influence score for 0038 FIG. 27 depicts an example of a view of cumulative that node has been determined or estimated. More specifi exposure of a post over a period of time on a Social network. cally, each Subject 102 may have an influence score indicating 0039 FIG.28 depicts an example of a view of social media the degree to which the subject’s opinion influences other content items displayed over a set of geographic locations on Subjects and/or a community of subjects, and each object 106 a map. may have an influence score indicating the collective opin 0040 FIG. 29 depicts an example of a flowchart of a ions of the plurality of subjects 102 citing the object. process to Support geo-tagging of social media content items 0046. In some embodiments, subjects 102 representing collected over social networks. any entities or sources that make citations may correspond to one or more of the following: DETAILED DESCRIPTION OF EMBODIMENTS 0047 Representations of a person, weblog, and entities 0041. The approach is illustrated by way of example and representing Internet authors or users of Social media not by way of limitation in the figures of the accompanying services including one or more of the following: blogs, drawings, in which like references indicate similar elements. Twitter(R), or reviews on Internet web sites: It should be noted that references to “an or 'one' or “some’ 0.048 Users of microblogging services such as Twit embodiment(s) in this disclosure are not necessarily to the terR; same embodiment, and Such references mean at least one. 0049. Users of social networks such as MySpace(R) or 0.042 A new approach is proposed that contemplates sys Facebook.(R), bloggers; tems and methods to identify geographic locations of all 0050 Reviewers, who provide expressions of opinion, Social media content items retrieved form a social network in reviews, or other information useful for the estimation of real time, wherein the geographic locations are physical loca influence. tions from which the Social media content items are origi nated or authored. If the latitude/longitude (geographic) coor 0051. In some embodiments, some subjects/authors 102 dinates of the content item are available, the geographic who create the citations 104 can be related to each other, for location of the social media content item can be identified a non-limiting example, via an influence network or commu using such geographic coordinates. For content items whose nity and influence scores can be assigned to the Subjects 102 geographic coordinates are not available, a historical archive based on their authorities in the influence network. of content items with high-confidence geographic locations is 0052. In some embodiments, objects 106 cited by the cita utilized to train a location classifier to predict geographic tions 104 may correspond to one or more of the following: locations of Such content items with high accuracy. Finally, Internet web sites, blogs, videos, books, films, music, image, the identified locations of the content items are confirmed to Video, documents, data files, objects for sale, objects that are be accurate and presented to a user together with the content reviewed or recommended or cited, Subjects/authors, natural items. or legal persons, citations, or any entities that are or may be 0043. As referred to hereinafter, a social media network or associated with a Uniform Resource Identifier (URI), or any social network, can be any publicly accessible web-based form of product or service or information of any means or platform or community that enables its users/members to form for which a representation has been made. post, share, communicate, and interact with each other. For 0053. In some embodiments, the links or edges 104 of the non-limiting examples, such social media network can be but citation diagram 100 represent different forms of association is not limited to, Facebook(R), Google+(R, Twitter(R), Linke between the subject nodes 102 and the object nodes 106, such dIn R., blogs, forums, or any other web-based communities. as citations 104 of objects 106 by subjects 102. For non 0044 As referred to hereinafter, a user's activities/content limiting examples, citations 104 can be created by authors items on a social media network include but are not limited to, citing targets at Some point of time and can be one of link, citations, Tweets, replies and/or re-tweets to the tweets, posts, description, keyword orphrase by a source/subject 102 point comments to other users’ posts, opinions (e.g., Likes), feeds, ing to a target (subject 102 or object 106). Here, citations may connections (e.g., add other user as friend), references, links include one or more of the expression of opinions on objects, to other websites or applications, or any other activities on the expressions of authors in the form of Tweets, blog posts, Social network. Such social content items are alternatively reviews of objects on Internet web sites WikipediaR) entries, referred to hereinafter as citations, Tweets, or posts. In con postings to Social media Such as Twitter R or Jaiku (R), postings trast to a typical web content, whose creation time may not to websites, postings in the form of reviews, recommenda always be clearly associated with the content, one unique tions, or any other form of citation made to mailing lists, characteristic of a content item on the Social network is that newsgroups, discussion forums, comments to websites or any there is an explicit time stamp associated with the content, other form of Internet publication. making it possible to establish a pattern of the user's activities 0054. In some embodiments, citations 104 can be made by over time on the social network. one subject 102 regarding an object 106, Such as a recom 0045 FIG. 1 depicts an example of a citation graph/dia mendation of a website, or a restaurant review, and can be gram 100 comprises a plurality of citations 104, each describ treated as representation an expression of opinion or descrip ing an opinion of the object by a source/subject 102. The tion. In some embodiments, citations 104 can be made by one nodes/entities in the citation diagram 100 are characterized Subject 102 regarding another Subject 102. Such as a recom into two categories, 1) Subjects 102 capable of having an mendation of one author by another, and can be treated as opinion or creating/making citations 104, in which expres representing an expression of trustworthiness. In some sion of Such opinion is explicit, expressed, implicit, or embodiments, citations 104 can be made by certain object imputed through any other technique; and 2) objects 106 cited 106 regarding other objects, wherein the object 106 is also a by citations 104, about which subjects 102 have opinions or Subject. US 2014/0040371 A1 Feb. 6, 2014
0055. In some embodiments, citation 104 can be described perform a query search over one or more social networks. As in the format of (Subject, citation description, object, times used hereinafter, keywords are the basic units for searches tamp, type). Citations 104 can be categorized into various and can be grouped into saved topics as shown in the non types based on the characteristics of subjects/authors 102. limiting example of FIG. 3. Social media content collection objects/targets 106 and citations 104 themselves. Citations engine 102 enables the user to enter multiple keywords by 104 can also reference other citations. The reference relation entering them as a comma-delimited list. For a non-limiting ship among citations is one of the data sources for discovering example, entering "egypt, Syria, libya, Sudan” in the search influence network. box will automatically input these four terms as separate 0056 FIG. 2 depicts an example of a system diagram to keywords for the search. When multiple keywords are Support interactive presentation and analysis of content over entered, social media content collection engine 102 does Social networks. Although the diagrams depict components search and keyword matching over the Social media content as functionally separate. Such depiction is merely for illustra utilizing OR operators. For a non-limiting example, a search tive purposes. It will be apparent that the components por query with keywords: jani 21, #feb 17, Egypt, Libya will trayed in this figure can be arbitrarily combined or divided match all results associated with itjan21, it feb 17, Egypt, OR into separate Software, firmware and/or hardware compo Libya. In some embodiments, Social media content collection nents. Furthermore, it will also be apparent that Such compo engine 102 also Supports Boolean search for content searches nents, regardless of how they are combined or divided, can to enable both OR and AND operators. execute on the same host or multiple hosts, and wherein the 0061. In some embodiments, social media content collec multiple hosts can be connected by one or more networks. tion engine 102 utilizes explicit first order literal matching of 0057. In the example of FIG. 2, the system 200 includes at keywords over the Social networks. Specifically, Social media least Social media content collection engine 102, Social media content collection engine 102 may search for keywords in a content analysis engine 104, and Social media geo tagging citation/Tweet's text field. If a Tweet is a native re-tweet, engine 106. As used herein, the term engine refers to soft then Social media content collection engine 102 searches in ware, firmware, hardware, or other component that is used to the citation/Tweet's retweeted status->text field. Here, effectuate a purpose. The engine will typically include Soft keyword matches of the Social content are case-insensitive. ware instructions that are stored in non-volatile memory (also For a non-limiting example, gadaffi will match gadafi or referred to as secondary memory). When the software Gadaffi or GADAFFI but will not match on kadaffi or instructions are executed, at least a Subset of the Software qadhafi or 'igadafi, and ilgadafficrimes will match instructions is loaded into memory (also referred to as pri itgadafficrimes or '#Gadafficrimes but will not match on mary memory) by a processor. The processor then executes gadafficrimes. the Software instructions in memory. The processor may be a 0062. In some embodiments, social media content collec shared processor, a dedicated processor, or a combination of tion engine 102 may remove punctuations determined as shared or dedicated processors. A typical program will extraneous when matching the keywords. Here, the punctua include calls to hardware components (such as I/O devices), tions to be ignored when matching keywords include but are which typically requires the execution of drivers. The drivers not limited to, the, to, and, on, in, of for, i, you, at, with, it, by, may or may not be considered part of the engine, but the this, your, from, that, my an, what, as, For a non-limiting distinction is not critical. example, if airplane' or airplane! appeared in the Tweets 0058. In the example of FIG. 2, each of the engines can run text as a standalone word or at the end of a tweet, then it would on one or more hosting devices (hosts). Here, a host can be a return as a match for airplane. computing device, a communication device, a storage device, 0063. In some embodiments, social media content collec or any electronic device capable of running a software com tion engine 102 enables matching based on commonly used ponent. For non-limiting examples, a computing device can citation conventions on Social networks. For a non-limiting be but is not limited to a laptop PC, a desktop PC, a tablet PC, example, Social media content collection engine 102 would an iPod(R), an iPhone(R), an iPad(R), Google's Android R. device, enable the user to match on citations/tweets about a stock by a PDA, or a server machine. A storage device can be but is not using the common Twitter R convention for referencing a limited to a hard disk drive, a flash memory drive, or any stock by inserting a dollar sign in front of the ticker symbol, portable storage device. A communication device can be but e.g., Tweets about Apple can be matched using the keyword is not limited to a mobile phone. Saapl which will match all tweets that contain the text 0059. In the example of FIG. 2, each of the engines has a Saapl or ‘SAAPL. communication interface (not shown), which is a Software component that enables the engines to communicate with 0064. In some embodiments, the user interface of the each other following certain communication protocols. Such social media content collection engine 102 further provides a as TCP/IP protocol, over one or more communication net plurality of search options via a search menu (shown as the works (not shown). Here, the communication networks can gear image to the left of keywords in the example of FIG. 3). be but are not limited to, internet, intranet, wide area network For non-limiting examples, such search options include but (WAN), local area network (LAN), wireless network, Blue are not limited to: tooth R., WiFi, and mobile communication network. The 0065 Newtopic, which clears the current list of keywords/ physical connections of the network and the communication search terms and allows the user to start a new search from protocols are well known to those of skill in the art. scratch. It is important to make Sure that the current topics (list Search of Contents over Social Network of keywords and parameters) are saved before proceeding to 0060. In the example of FIG. 2, social media content col the new topic. lection engine 102 searches for and collects Social media 0.066 Enable all, which turns on all keywords listed for the content items (e.g., citations, Tweets, or posts) by enabling a search whether they are currently enabled (not grayed out) or user to enter one or more keywords via a user interface to not. US 2014/0040371 A1 Feb. 6, 2014
0067 Revert topic, which refreshes the search results with Filtering of Social Media Content the keywords and parameters from the current topic. 0076. In the example of FIG. 2, social media content col 0068 Share topic, which shares the list of keywords and lection engine 102 enables a user to refine search results by parameters easily with others by cutting and pasting the URL the a plurality of parameters, which include but are not limited into an email or an instant message. to dates, locations, languages, sentiment, Source, and influ 0069. In some embodiments, the social media content col ence of the citations/Tweets collected as shown by the lection engine 102 provides at least two options for the dis example depicted in FIG. 6. When multiple filters are playing keywords in the search result: selected, they will be applied by social media content collec tion engine 102 based on the AND operator. For a non 0070) Enabled, which displays the one keyword or mul limiting example, selecting a location filter on Libya, Syria, tiple keywords selected in the analysis of the search result. Lebanon and language filter on English will match all 0071 Isolated, which automatically turns off all the key results located in Libya, Syria, OR Lebanon AND in the words other than the one selected in the analysis of the search English language. result. 0077. In some embodiments, social media content collec tion engine 102 enables the user to restrict the search results Exporting and Sharing of Social Media Content based on dates/timestamps of the citation. For a non-limiting example, the default selection of time range can be last 24 0072. In some embodiments, the social media content col hours, which can be changed to any of the following: last lection engine 102 enables the user to save user-defined sets hour, last 24 hours, last 7 days, last 30 days, last 90 days, last of keywords and report parameters that define a search as a 180 days, or a specific date range as specified by the user as saved topic/search. Saved topics can be used as logical group shown by the example depicted in FIG. 7. ings of terms/keywords commonly associated with a particu 0078. In some embodiments, social media content collec lar country or event (e.g., Hegypt, #mubarak, #muslimbroth tion engine 102 enables the user to filter the search results erhood, Hijan25, (aegyptocracy). Such saved topic or search based on the originating locations of the citations/posts/ allows users to save keywords and parameters so they can be Tweets. Here, the filtering location can be specified at the used again as shown in the example depicted in FIG. 4. country, State, county, or city level. Additionally, the filtering 0073. In some embodiments, social media content collec location can be specified by latitude and longitude coordi tion engine 102 provides a saved search dropdown menu, nates as shown by the example depicted in FIG. 8. which allows the user to easily find and retrieve previously 0079. In some embodiments, social media content collec saved topics. If there are a lot of saved searches, the user can tion engine 102 searches and returns search results for Social enterparts of the saved search name in a searchbox to find the media content in any language regardless of character set. specified search topic as shown in the example depicted in Since Social media content collection engine 102 matches the FIG.S. content items based on literal keywords, the user can enter any word from a foreign language and Social media content 0.074. In some embodiments, social media content collec collection engine 102 will return exact matches for the words tion engine 102 enables a user to download a saved topic/ entered. In addition, Social media content collection engine search and the corresponding search results from the topic to 102 uses various methods of language morphology (e.g., a specific file/date format (e.g., CSV format) by clicking the tokenization) to isolate search results to just the language Export button on the user interface. In addition, social media specified for a specific set of languages, which include but are content collection engine 102 may also provide an Applica not limited to English, Japanese, Korean, Chinese, Arabic, tion Programming Interface (API) URL for users who want to Farsi and Russian as shown by the example depicted in FIG. access the Secure Reporting API to programmatically retrieve 9. data. All citations/Tweets from the search query can be down 0080. In some embodiments, social media content collec loaded in batch mode, including those 'significant posts'. tion engine 102 adopts various language detection and pro which are tweets that have links or tweets that have been cessing techniques to filter the search results by language, re-tweeted. wherein the language detection techniques includebut are not 0075. In some embodiments, social media content collec limited to, tokenization, domain-specific handling, Stemming tion engine 102 enables a user to copy a topic by clicking the and lemmatization. Here, the tokenization of the search “Save As ... button and choosing “Create a new Topic' to results is language dependent. Specifically, whitespace and save a copy of the existing topic under a new name. Social punctuation are delimited for European languages, Japanese media content collection engine 102 further enables a user to is tokenized using grammatical hints to guess word bound share a topic with another user by clicking the gear icon next aries, and other Asian languages are tokenized using overlap to the list of keywords (as shown in FIG. 3) and choosing the ping n-grams. As referred to hereinafter, an n-gram is a con Share topic menu option. Social media content collection tiguous sequence of n items/words from a given sequence of engine 102 will generate a unique topic URL that the user can text or speech, which can be used by a probabilistic model for copy to share with another user. For a non-limiting example, predicting the next item in Such a sequence. the topic URL can be in the format: https://"SOCIAL 0081. In some embodiments, social media content collec ANALYSIS SYSTEM.topsy.com/share/view?id=DXXX, tion engine 102 uses character set processing as a first pass where XXX is a unique topic id. Any user on the same Social through character sets (e.g., Chinese, Japanese, Korean), media contentanalysis system/platform can view atopic from while statistical models can be used to refine other languages another user as long as the user is given a valid Topic URL. (English, French, German, Turkish, Spanish, Portuguese, Please note that social media content collection engine 102 Russian), and n-grams be used for Arabic and Farsi. In some keeps all topic URLS private and requires a system account embodiments, domain-specific handling is utilized to identify for another user to login to view the topic. and handle short Strings and domain-specific features such as US 2014/0040371 A1 Feb. 6, 2014
#hashtags, RT (a replys for search results from Social net social media source IDs (e.g., “barackobama') will fre works such as Twitter R. Stemming and lemmatization fea quently have high influence because they are mentioned by tures are available for English and Russian languages. many influential users, including news organization. Like 0082 In some embodiments, social media content collec wise many celebrities (e.g., justinbieber') have high influ tion engine 102 utilizes a user's historical comments/posts/ ence since they are frequently mentioned by other influential citations to improve accuracy for language detection for users. In some embodiments, social media content collection search results. If the user is consistently identified as a user of engine 102 utilizes a decay factor, so that an account of a user one specific language upon examining his/her historical com which is inactive—and which therefore no other user is men ments, future comments from that user will be tagged with tioning will fall to the bottom of the influence ranking, as that specific language, which largely eliminates false nega will an account from spammers or celebrities who do not post tives for such user. things that other influential users find interesting. 0083. In some embodiments, social media content collec 0086. In some embodiments, social media content collec tion engine 102 filters search results by whether the user/ tion engine 102 adopts iterative influence calculation to author sentiment of the content item collected is positive, handle the apparent circularity of the influence level (i.e., that neutral, or negative based on analysis of the posted English an individual gains influence by receiving attention from text as shown by the example depicted in FIG. 10. Specifi other influential individuals) by measuring centrality of an cally, Social media content collection engine 102 uses a attention graph/diagram. As shown in the example depicted in curated dictionary of sentiment weighted words and phrases FIG. 12, every author is a node on the directed attention to fine tune its sentiment detection techniques to handle con diagram, and attention (mentions, reposts) are edges. Cen tent from a specific Social network, such as Twitters(R unique trality on this attention diagram measures the likelihood of a 140 character limits and “twitterisms’. By combining some person receiving attention from any random point on the English grammar rules to this, social media content collection diagram. In the example of FIG. 12, Author F has reposted or engine 102 is able to accurately fine tune results in relatively mentioned Authors C.D. and G so there are outgoing line high accuracy rates, with results typically garnering a 70% edges from F to these authors. Author D has received the most agreement rate with manually reviewed content. Social media attention and is likely to be influential, especially if most of content collection engine 102 is further able to identify and the authors mentioning or reposting Author Dare influential. ignore entities with misleading names (e.g. Angry Birds) and applying Stemming and lemmatization to expand the senti Dashboard Presentation of Social Media Content ment dictionary scope. Here, the curated dictionary of senti I0087. In the example of FIG. 2, social media content ment weighted words and phrases can grow organically based analysis engine 104 presents a dashboard that shows a quick on real world data as more and more search results are gen Snapshot of what is important for the given search keywords erated and grammar rules found to be significant in helping to and parameters selected by the user. If nothing is selected the determine sentiment are included. For a non-limiting default is to show everything trending on the Social network example, the use of the word “not” before a word is used as a right now or for the past 24 hours. The content Snapshots negativity rule. In Addition, since Stemming can introduce presented within the dashboard include one more of: Activity, errors in categorization of sentiment (example, the root by Top Posts, Top Links, and Top Media and the user may navi itself could have negative sentiment but root--suffix could gate to the dedicated view of a specific Snapshot of the content have positive sentiment). Such stemming errors are handled by clicking on the title of the content (e.g., clicking on Top on a case by case basis by adding the improper sentiment Posts takes the user to the Trending Tab with Posts selected). categorization due to stemming as exceptions to the dictio Specifically, nary. I0088 Activity snapshot shows the number of mentions for 0084. In some embodiments, social media content collec the top five (if more than five keywords are entered) most tion engine 102 filters search results to those authored by active keywords entered in the search box as shown by the users determined to be influential only as shown by the example depicted in FIG. 13. As shown in FIG. 13, the data example depicted in in FIG. 11. Here, influence level of an displayed represent the number of total mentions for each author measures the degree to which the author's citations/ keyword within the time range selected, as well as the most Tweets/posts are likely to get attention from (e.g., actively related terms to the target keyword(s) selected within the time cited by) other users, wherein various measures of the atten range specified. tion Such as reposts and replies can be used. The influence I0089 Top Posts snapshot shows the top four significant level can be from a scale 0 to 10 and the influence filter will posts for the keywords entered along with their number of only return results from users who are “highly influential mentions. The posts are ranked by relevance so the most (10) or “influential” (9). Such influence level can be deter important posts are displayed as shown by the example mined based on a log scale so influence of a user has a very depicted in FIG. 14. If a number of keywords are entered, then skewed distribution with the “average' influence level being the posts are compared against each other to determine which set as 0. The influence measures are resistant to spamming, posts from which keywords are displayed. The social media since an author cannot raise his or her influence just by having content analysis engine 104 attempts to display at least one lots of followers, or by having a large value of some other post from each keyword if there are less than four keywords easily inflatable metric. They must be other authors. entered. 0085. In some embodiments, social media content collec 0090 Top Links snapshot shows the top six trending links tion engine 102 calculates the influence level of a user tran for the keywords entered along with their number of mentions sitively, i.e., the users influence level is higher if he/she as shown by the example depicted in FIG. 15. receives attention from other people with influence than if the 0091 Top Media snapshot shows the top trending videos user receives attention from users without influence. For a and photos for the keywords entered as shown by the example non-limiting example, the politicians as identified by their depicted in FIG. 16. US 2014/0040371 A1 Feb. 6, 2014
Activity History Over Social Network to quickly and easily change the range to see the time frame that is relevant to his/her analysis as shown by the example 0092. In some embodiments, social media content analy sis engine 104 provides activity history view that displays the depicted in FIG. 19. 0101. In some embodiments, social media content analy volume of mentions for a set of keywords over a period of sis engine 104 enables the user to select and view the Top time. Social media content analysis engine 104 provides the Posts with a specific time range selected. If a specific point on user with the ability to select the start and end dates for the activity diagram is selected, then the Top Posts are from displaying mention metrics within the view/report. It also just that date and keyword selected. For a non-limiting enables the user to specify the time windows to display, example, if the top peak of the dark green line was selected, including by month, week, hour, and minute. Such a view/ the top posts for iiNBA at 6PM will be shown by the example report is useful for examining historical events and identify depicted in FIG. 20. Here, the Tops Posts shows the top ing patterns. For non-limiting examples, such report can be significant posts (posts that contain a link or are re-posted) for used to: that specific day and do not necessarily show all the posts for 0093 Track the number of mentions of the leading US a given day. Presidential contenders (Obama, Romney, Gingrich, and 0102. In some embodiments, social media content analy Santorum) over the past six months. sis engine 104 a list of the most recent trending metrics for the 0094 Track the number of negative sentiment mentions specified saved search group or keywords/terms entered/ for the President using the following keywords: Obama, mentioned by social media content items. Each term will iobama, President Obama, (abarackobama, and (a)white include the following metrics: mentions, percent influence, house based on the following locations: in Egypt, Libya, momentum, Velocity, peak period as shown by the example Syria, Lebanon, Israel, and Iraq. depicted in FIG. 21. Users can view these metrics so they can 0.095 Track the number mentions in Chinese of Foxconn quickly identify what terms have the highest mention Volume, in China, Hong Kong, Taiwan. are trending the most via momentum, or are peaking most 0096. Track the number of hashtags representing Syrian recently via peak period metrics. cities over time, isolating the mention activity to Arabic lan gllage. Discovery of Related Terms 0097. In some embodiments, social media content analy sis engine 104 makes the activity history data available for 0103) In some embodiments, social media content analy presentation in real time on a rolling basis. Specifically, sis engine 104 enables the dynamic discovery of new terms minute metrics are available for the last 6-8 hours on a rolling that are related to existing known keywords submitted for a basis, hour metrics are available for last 30 days on a rolling query search over a social network. Such discovery presents basis, and daily metrics are available at least 6 months back. the user with a list of top keywords (e.g., individual words, 0098. In some embodiments, social media content analy hashtags or phrases in any language) related to and/or co sis engine 104 allows the user to enable and disable the occurring the one(s) entered by the user and trending (cur keywords and their associated lines on the figure by clicking rently or over a period of time) over the social network based on the keywords below the figure as shown by the example on various measurements that that measure the trending char depicted in FIG. 17. This is a very useful feature when a acteristics of the terms in the Social media content items number of keywords are graphed, with a few “flooding out collected over a period of time. For non-limiting examples, the others due to high Volume. Removing these higher Volume Such measurements include but are not limited to mentions, keywords by simply clicking on them enables the user to influence, Velocity, peak, and momentum, which can be cal "peel back layers of smaller volume lines to identify what culated by the Social media content analysis engine 104 based activity may be important over time. on all citations/tweets/posts containing the search keywords 0099. In some embodiments, social media content analy AND the discovered related terms. This list of related words/ sis engine 104 supports Share of Voice (SOV) analysis, which terms enables the user to see what terms are related to known is used to measure the relative change in mention activity for terms submitted, the strength of their relationships and the a given group of keywords over time as shown by the example extent to which each of the related terms are trending within depicted in FIG. 18. SOV analysis calculates the total number the time range selected. For a non-limiting example, different of mentions for a keyword and divides the number of men trending terms co-occurring and related to the search term tions for a keyword by the Summed amount of mentions for “Republican” (e.g., Gingrich, Romney, Ryan, etc.) can be the group of keywords being analyzed so the relative percent discovered over different phrases of the 2012 presidential age of each keyword's mentions over time can be analyzed. campaign cycle, which can then be used to search for most The metrics used in a SOV analysis could also be scoped for relevant social media content items within the relevant time a specific language, social data Source or geographic area. periods. This is a useful technique for measuring the relative impor 0104 For non-limiting examples, the related terms dis tance of Something being mentioned on the Social web over covered by Social media content analysis engine 104 enables time within a given category of related keywords or phrases the user to: and other parameters. 0105 Determine the top trending keywords/events/ 0100. In some embodiments, social media content analy people?hashtags that the user does not know about for a sis engine 104 enables the user to select a time slice window known list of keywords. for the date range requested whether that is by minutes, hours, 0106 Discover what terms are most highly correlated to day, week, or month and to Zoom in and out on a specific keywords now, 6 weeks ago, or even 6 months ago. The region of the activity diagram by clicking a region and then related discovery terms are determined based on the time holding down the click until identified the region to Zoom into range selected so analysis can be done to see how terms has been selected (click & drag to select). This allows the user change over time. US 2014/0040371 A1 Feb. 6, 2014
0107 Identify keywords related to single known term, engine 102. Weight is given to the terms whose absolute rate building awareness based upon the knowledge gleaned from of co-occurrence with the query is larger than others. discovering new terms. 0115 Intentional, where a bonus or weight is given to 0108 Quantify what terms are most related, and have the hashtags because they suggest an intent to query. highest Volume or most recent peaks based upon analysis of 0116. In some embodiments, social media content analy the metrics. sis engine 104 also discovers and/or sorts the related terms 0109. In some embodiments, social media content analy based on one or more of momentum, Velocity, peak and sis engine 104 discovers the related terms by examining a influential metrics in addition to correlation scores and men historical archive of recent tweets/posts retrieved from the tions (e.g., total number of mentions/re-tweets for this post, Social network for top trending terms co-occurring with the link, image or video over its lifetime) for each of the related submitted keywords before searching over the social net terms. The following metrics are based on the timeframe set work. The discovered related terms can then be used together by the user in the search parameters and are calculated off of with the keyword(s) submitted by the user to search for the a census-based post index for all posts: relevant content items in the Social media content stream 0117 Momentum, which measures the combined popu retrieved continuously in real time from the social media larity of a term and the speed at which that popularity is network via a social media source fire hose. Alternatively, increasing. A high score indicates that there have been more Social media content analysis engine 104 may dynamically frequent recent citations/posts relative to historical post activ discover the related terms by examining the Social media ity. Terms with high momentum scores typically have high content stream in real time as they are being retrieved and levels of post Volume. For a non-limiting example, momen apply the related terms discovered to search for relevant tum for the past 24 hours can be calculated as: Social media content items together with the user-Submitted momentum-sum of (h/24*count of h), where h is the hour, keyword(s). from 1 to 24, 24 being the most recent hour. 0110. In some embodiments, social media content analy 0118 Velocity, which solely measures the speed at which sis engine 104 discovers the related terms via a significant a terms popularity is increasing, independent of the terms post index, which includes citations/posts that contain a link overall popularity. Velocity numbers can be in the range of or a re-post to another content item. Social media content 0-100. If the time window is 24 hours, then 100 means that all analysis engine 104 then applies a weighted frequency analy volume over that time period selected happened within the sis to the significant posts containing the Submitted keywords past hour. The difference between momentum and velocity is and the related terms to discover the related terms within the that Velocity only measures speed while momentum mea date range selected. Sures both speed and popularity (Volume of mentions). For a non-limiting example, Velocity over the past hour can be 0111. In some embodiments, social media content analy calculated as: velocity=(100*momentum)/mass, where h is sis engine 104 discovers and/or sorts the list of related terms the hour, from 1 to 24, 24 being the most recent hour, and based on a combination of one or more of: mass is Sum of count of h- i.e. just the total count over the 0112 Unexpected, where weight is given to the terms that 24 hour period. are uncommon in the general search, which means the daily 0119 Peak, which indicates the time period that had the scale document frequency is low, i.e. a result term that has not highest number of contentitems containing the terms over the been mentioned a lot in the last few days. For a non-limiting time period selected. The unit is calculated based on the date example, if both “foreign ministers' and “vehicles' are range selected, including 24 hours (unit of measure is hours), appearing for query “Syria' and have equal levels of co 7 days (unit of measure is days), 30 days (unit of measure is occurrence with the query (same number of tweets in last few days), 90 days (unit of measure 180 days (unit of measure is hours containing both “syria' and “foreign ministers' as the weeks), and specific date range, where unit of measure is number of tweets containing both “syria' and “vehicles”), calculated based on the time frame that is entered. If the then “foreign ministers' is likely to rank higher because specified date range is less than a year, then the above unit “vehicles' is a more common term and is used more often in measurements are utilized. If the date range is longer than a other contexts (as measured over the last few days). year then the peak period is based on a time slice out of 52 0113 Contemporaneous, where weight is given to terms across the time period. whose rate of co-occurrence with the keywords submitted has 0120 Influence, which measures the total number of influ increased significantly in a short period of time. The discov ential mentions/retweets of a content item (e.g., post, link, ered terms become available in real-time and it is possible to image or video) containing the terms over the lifetime of the query historical time intervals. The metrics used to track content item. Social media content analysis engine 104 increases for the terms over time is gathered in a counting counts influential mentions for a link as the total number of bloom filter fed by search index of significant tweets/posts. tweets that have contained the link from influential users. For For each term and term-pair, Social media content analysis a non-limiting example, an influential count of 5 could mean engine 104 keeps an estimate of the frequency on both an that there were 5 different tweets from 5 different influential hourly and daily scale. From this the Social media content users or 5tweets containing the link from one influential user. analysis engine 104 computes an estimate of the Velocity and 0.121. In some embodiments, once the terms related to the momentum whenever the Velocity and momentum exceed set of keywords have been discovered, social media content certain thresholds it emits a term pair. It should be possible to analysis engine 104 utilizes both to search the social network identify the related terms with spikes or rises in the standard for the content items (citations, tweets, comments, posts, etc.) metrics containing all or most of the keywords plus the related terms. 0114 Meaningful, where phrases are filtered for quality For a non-limiting example, the top posts found by search via against Wikipedia, Freebase, and other open databases, as the target/submitted search term and the discovered related well as the query logs of the Social media content collection term as shown by the example depicted in FIG. 22. In some US 2014/0040371 A1 Feb. 6, 2014
embodiments, where there may not be a comment that has all I0127 View what news stories are trending about keyword the terms, social media content analysis engine 104 attempts “Syria”. to determines the top content items that contains as much of I0128 View what news stories on wsj.com and nytimes. the keywords, including the related term as possible. Conse com have the highest Volume of mentions or percentage of quently, one post/content item may appear for every related influencers over the past 24 hours. term in the search results. I0129. Compare which news stories/links have the highest momentum between the New York Times (nytimes.com) and Trending the Washington Post (washingtonpost.com). 0122. In some embodiments, social media content analy 0.130 Isolate what links are trending the most within a sis engine 104 presents the top trending results for posts, country by only selecting country and not specifying any links, photos, and videos sorted by one or more of relevance, thing else. date, momentum, Velocity, and peak based upon the time I0131. In some embodiments, social media content analy frame selected. It is important to note that all these metrics are sis engine 104 presents the top trending media (photos and always scopes to the post, link, photo or video displayed—not videos) related to the keywords and parameters entered. The the aggregate number of posts for a given keyword. The Social results presented can be sorted by one or more of relevance, media content analysis engine 104 identifies the most signifi date, momentum, Velocity, and peak as shown by the example cant posts which were mentioned within the time range depicted in FIG. 26. Displayed along with the top photo, selected, with variations in the metrics presented that are which can be shared on the social network (e.g., Twitter(R) important to note. In addition, for all the time ranges from from a variety of photo sharing sites (e.g., twitpic, yfrog, X-date to present (e.g., past 24hours, past 7 days), the mention instagram, twimg), are the number of mentions containing and influential mentions are calculated based on the number the photo link, number of influential people that posted the of all-time mentions. If a specific time slice is selected (e.g., link, and the momentum, Velocity, and peak score. In some Jan. 1, 2012 to Jan. 31, 2012) then the mention and influence embodiments, a spark line is displayed in order to quickly metrics are also scoped to all time and not to just the time determine what photo is trending or stale. The view of trend frame specified. ing photos is very useful for identifying photos associated 0123. In some embodiments, social media content analy with events as they unfold. Such view can be used to find sis engine 104 presents the trending top posts for the key photos from individuals on the ground before media outlets words and parameters specified, where the view displays the pick them up. Users can also isolate what photos are trending actual post, along with the author of the post, a timestamp of the most within a country by only selecting country and not when the post was originally communicated, and the corre specifying anything else. sponding mention, influential (number of influential men 0.132. In some embodiments, social media content analy tions), momentum, Velocity, and peak metrics. In addition, sis engine 104 presents the top trending videos related to the the profile information of the user on the Social network (e.g., keywords and parameters entered. The results presented can Twitter R is displayed (name, link, bio, latest post, number of be sorted by one or more of relevance, date, momentum, posts, number they are following, and number of followers) Velocity, and peak. Displayed along with the top video, which by highlighting the picture associated with the user's login is shared on the social network (e.g., TwitterR) from a variety name on the social network. The user is also enabled to click of video sharing sites are the number of mentions containing on the arrows on the right side of the spark line diagram for the video link, number of influential people that posted it, and each post from the view depicted in FIG. 23, which displays the momentum, Velocity, and peak score. In some embodi the overall activity of that specific post for the lifetime of the ments, a spark line is displayed to quickly determine what post as shown in the example depicted in FIG. 24. video is taking off (i.e., trending) or stale. The view of trend 0124. In some embodiments, social media content analy ing videos is very useful for identifying videos associated sis engine 104 presents the trending links, where the view with events as they unfold. Such view can be used to find displays the most popular links matching any set of keywords, videos from individuals on the ground before media outlets including domains. By specifying only domains as keywords pick them up. Users can also isolate what videos are trending (e.g., "nytimes.com'), the trending links view returns the the most within a country by only selecting country and not most popular links on a specific domain/website (e.g., wash specifying anything else. ingtonpost.com, espn.com) or across the multiple domains I0133. In some embodiments, social media content analy entered. For each domain specified, social media content sis engine 104 presents the trending terms report, which pro analysis engine 104 will display one or more of the following vides a list of the most recent trending metrics for the speci metrics: mentions, percent influence, momentum, Velocity, fied saved search group or keywords entered. Each term may peak period. include one or more of the following metrics: mentions, per 0.125. In some embodiments, social media content analy cent influence, momentum, Velocity, peak period. Users can sis engine 104 enables the user to input multiple domains for view these metrics to quickly identify what terms have the domain analysis in order to quickly identify what links these highest mention Volume, are trending the most via momen domains have the highest mention Volume, momentum, tum, or are peaking most recently via peak period metrics. Velocity or are peaking most recently via peak period metrics as shown in the example depicted in FIG. 25. Such analysis Exposure identifies what articles/links are most popular on any domain consumers are referencing within the Social network (e.g., I0134. In some embodiments, social media content analy TwitterR). For non-limiting examples, such analysis can be sis engine 104 presents the a cumulative exposure view of the utilized to: search results, which returns the gross cumulative exposure 0126 Analyze which stories have just broken and are the for posts containing a set of keywords overtime. This analysis most popular over the past 24 hours on aljazeera.com. is useful to measure the gross exposure over time from posts US 2014/0040371 A1 Feb. 6, 2014
matching a target set of keywords. For non-limiting from the public list of social media sources lists and their examples, such cumulative exposure view can be used to: respective verified accounts and grown organically on an 0135 View the number of cumulative gross impressions ongoing basis. of a specific post, such as a speech delivered by President 0144. In some embodiments, social media content analy Obama’s Middle East speech (timespeech) in Libya, Syria, sis engine 104 may review the user's profile and historical and Egypt for the 24 hours after he delivered the speech. post information to intelligently identify media/news sources the user belongs to. Some of the attributes and features of the 0.136 View the cumulative negative sentiment exposure of users information being reviewed by Social media content a hot topic with certain time frame, such as idebtcrisis for the analysis engine 104 include but are not limited to: first week of September 2011. 0145 Total number of posts 0.137 View the cumulative exposure of the keywords 0146 Total number of reposts referring to a specific person over a period of time, such as 0147 Percentage of posts that have links Medvedev, ilmedvedev, and (amedvedevrussia in Russian in 0.148 Percentage of posts that are (a replies the US and Russia over the past 30 days. 0.149 Total number of distinct domains from links 0138 Identify "tipping points' in when gross exposure posted significantly increased for a given set of terms over time. 0.150 Average daily post count 0.139. In some embodiments, social media content analy 0151. Similarities to other media accounts sis engine 104 calculates the cumulative exposure by Sum 0152 Profile URL matches a media site ming the follower counts of all the authors of the posts that 0153. Profile name of user matches a media name or a match the keywords being queried. This calculation returns real human name overall gross exposure (vs. unduplicated net exposure) so multiple posts from the same author or authors with common Geography followers may result in audience duplication as shown by the 0154) In some embodiments, social media content analy example depicted in FIG. 27. sis engine 104 Supports geographic analysis, which returns/ 0140. In some embodiments, social media content analy presents a view/report on at least Some of the Social media sis engine 104 displays top significant posts in the cumulative content items (social mentions) with a set of known geo exposure view for the time range selected in the search param graphic locations over a period of time as shown by the eters. If a specific point on the exposure view is selected then example depicted in FIG. 28. Here, the geographic locations the top posts are from just that date and keyword selected. For refer to places where posts are originating from, and can be a non-limiting example, in the example depicted in FIG. 27, if defined by city, county, State, and countries. For non-US the dot on the line for the date 2/21 is selected then the top countries, “state' and “county’ correspond to administrative significant posts for Syria will be shown on that date. division levels. This report can be displayed on a world map with shading indicating the relative Volume of mentions at Cross Network ID their geographic locations on the map, wherein the world map can be Zoomed in to focus on the Social mentions at a region 0141. In some embodiments, social media content analy or a country and enables the user to drill-down to see the sis engine 104 Supports cross network identification to iden Social mentions at country, state, county, or city level. In some tify an author and to view the content produced by the same embodiments, the geographic analysis report shows country user across different social networks, such as between Twit level metrics at a high confidence and coverage rates. For a ter(R) and Blogs, or a review site and a chat site analysts. non-limiting example, a confidence rate of 90% means that Specifically, social media content analysis engine 104 com 90% of posts that are geo-tagged at the country level are pares the user profile photos and/or content of the posts from correct based on validation methods. different Sources of Social media content and analyzes if the 0.155. In some embodiments, social media content analy author is the same on those sources. If the same author is sis engine 104 shades the world map based on a polynomial identified, social media content analysis engine 104 may function that colors the map by default based on the raw assign a common cross network identification to the user. Volume of mentions per geographic location. If the Activity Social media content analysis engine 104 may further present table is re-sorted by "96 Activity”, then the world map is the user's posts over the different social media sources/social refreshed and shaded based on the relative percentage activity networks side-by-side on the same display in Such way to for each country. When the shaded location (the one selected enable a viewer to easily toggle between the different social as part of the report parameters) is rolled over, the volume networks to compare the posts by the same user. metrics and percent activity are displayed. The table below the map allows the user to seemention and percent metrics for Media Identification each geographic area. Here, the “96 Activity” metrics are defined as the mentions matching the entered keywords 0142. In some embodiments, social media content analy divided by total overall mentions for the geographic area. In sis engine 104 Supports media identification to classify indi Some embodiments, social media content analysis engine 104 vidual authors of Social media content items from commer may calculate the “96 Activity” metric by taking the total cial and news sources. By filtering out commercial and news posts for the keywords entered divided by the total number of Sources, social media content analysis engine 104 is able to all posts for that country, basically calculating a share of voice generates reports focused on individuals "on the ground'. percentage. For a non-limiting example, a 3.1% activity 0143. In some embodiments, social media content analy means that 3.1% of tweets found for that country contain the sis engine 104 uses a combination of a whitelist and a trained keywords entered during the timeframe specified. In some classifier to assign users as a media or non-media type. For a embodiments, social media content analysis engine 104 non-limiting example, the whitelist can initially be derived enables the user to display metrics by specifying either lati US 2014/0040371 A1 Feb. 6, 2014
tude/longitude or not, in which case metrics will be calculated 0163 Information about the software client used to post based upon the systems inferred geo location. the message on the Social media site (e.g. a particular mobile application for TwitterR). GeoTagging Methodology 0164 Content analysis, which parses the content within the post to identify locations within the content. Statistics can 0156. In the example of FIG. 2, social media geo tagging be applied to this data to uncover potential geo-location of engine 106 identifies and marks each Social media content users. Indirect content analysis includes, e.g., URLs or refer item with proper geographic location (geo location or geo ences to entities (including websites) that are known to be tag) from which Such content item is authored. In some associated with specific locations. The knowledge of the loca embodiments, social media geo tagging engine 106 is able to tion associations of Such entities may either be set explicitly identify geo-location of a social media content item using the (e.g., a local newspaper is explicitly associated to the city in latitude/longitude (lat/long) coordinates of the content item which it is published; the Empire State building is explicitly when the user/author of the content item opts in to share the associated to New York city) or such entity location associa GPS location of the digital device where the content item is tion may be inferred through a variety of methods including originated. Lat/long is highly accurate for identifying (i.e., the methods described here for associating location to posts identifies with high confidence value) where the user is when and users/authors of posts. he/she communicates via a mobile device but it may have 0.165 Geo-located hashtags for events in the post, where very sparse coverage (generally 1-3% of the posts) depending trending hashtags of known events are identified and associ on the query parameters used. Here confidence value is ated to the geographic location of where events occur for the expressed as the probability that a post came from a specified events time periods (e.g., a conference in NYC is trending location. In addition, Social media geo tagging engine 106 and people are posting about it using the hash-tag). Citations/ provides geo trace scores to help identify the relative weight Tweets containing hashtags of known events and tweeted of the geo adaptations/features used to identify the location. within the timeframe of the events time periods will be O157. In some embodiments, social media geo tagging associated to the events location. engine 106 may identify geo-location of a content item from 0166 In some embodiments, social media geo tagging the profile information of the author/user of the content item, engine 106 uses the high-confidence geolocation information wherein the user's profile contains the user's self-described in posts having Such information as anchors to identify geo geographic location. The data point in the user's profile iden graphic locations of other content items whose geographic tifies where the user may be (not where they are communi locations (e.g., geo-coordinates) are not available, with rela cating from) with low confidence (because the information is tively high level of confidence, to increase geographic loca self-described by the user him/herself) but with relatively tion coverage of the Social media content items significantly. high coverage (50-70%). Social media geo tagging engine Specifically, an archive of historical content items/posts with 106 determines that the location identified in the user's profile high-confidence geographic coordinate data can be used as a is “valid if the user with that location is generally telling the training set to train a customized probabilistic location clas truth (e.g. people who claim to live in Antarctica are generally sifier. Once trained, the location classifier can then be used to not telling the truth). predict the actual geographic locations of the content items 0158. In some embodiments, social media geo tagging without geo-coordinates with high accuracy. engine 106 may utilize one or more of the following for 0.167 During the training process, Social media geo tag geo-location identification in addition to use of lat/long coor ging engine 106 reversely geocodes the latitude/longitude dinates and user profile: coordinates of each post in the training set using an internal 0159 Language used in the post, which can be utilized to lookup table. Forgeo-tagged posts in the United States, social strengthen the confidence when used in combination with media geo tagging engine 106 assigns the location based on other methods for geo-identification. the lat/long point being found within a defined polygon, asso 0160 Exif (Exchangeable Image File Format) photo ciating each content item in the training set with the 4-tuple metadata of the post, which contains Lat/long data embedded be originated from San Francisco ( What is claimed is: tomized probabilistic location classifier to predict geo 1. A system, comprising: graphic locations of the content items which geographic a social media content collection engine, which in opera locations are not available. tion, continuously retrieves all social media content 10. The system of claim 9, wherein: items from a social network in real time, wherein the the Social media geo tagging engine assigns geographic retrieved social media content items do not have IP location to a content item in the training set in the form address identifiers to identify their geographic locations; of city, state, or country by reverse geocoding of the a social media geo tagging engine, which in operation, latitude/longitude coordinates of the content item. continuously identifies and marks each of the Social 11. The system of claim 9, wherein: media content items retrieved in real time with its the location classifier recognizes and extracts a set of fea proper geographic location from which Such content tures related to geographical location from each of the item is authored or originated; posts in the training set and calculates an observation set confirms the identified location of a content item by of the extracted features as the cross-product of the comparing the identified location to locations of prior location vector and feature set, yielding data of the author of the content item, time stamp of the set and calculates an observation set of the extracted content item, content analysis of the content item, and features as the cross-product of the location vector and geo-located hashtags for events in the content item. feature set, yielding