{"id":435792,"date":"2019-02-14T11:00:22","date_gmt":"2019-02-14T19:00:22","guid":{"rendered":"http:\/\/www.esri.com\/arcgis-blog\/?post_type=blog&p=435792"},"modified":"2019-02-14T21:22:59","modified_gmt":"2019-02-15T05:22:59","slug":"centrality-measures-and-link-analysis-insights-for-arcgis-use-link-charts-search-for-okcupid-matches","status":"publish","type":"blog","link":"https:\/\/www.esri.com\/arcgis-blog\/products\/insights\/analytics\/centrality-measures-and-link-analysis-insights-for-arcgis-use-link-charts-search-for-okcupid-matches","title":{"rendered":"Love and Learn Link Analysis Using Insights for ArcGIS"},"author":6671,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","format":"standard","meta":{"_acf_changed":false,"_searchwp_excluded":""},"categories":[23341],"tags":[327672,43641,327682,327662,36431],"industry":[],"product":[36801],"class_list":["post-435792","blog","type-blog","status-publish","format-standard","hentry","category-analytics","tag-dating","tag-link-analysis","tag-okcupid","tag-recommender-systems","tag-social-media","product-insights"],"acf":{"short_description":"Use Insights for ArcGIS to find hidden links and relationships in your data. Offer intelligent recommendations using centrality measures.","flexible_content":[{"acf_fc_layout":"content","content":"

Have you ever cracked a little smile, when your smart phone seems to share your own intuition or recommend something you adore?\u00a0 Let’s examine these analytical moments through a\u00a0link analysis<\/em>\u00a0using Insights for ArcGIS<\/a>.<\/p>\n

To do this, our team thought it would be fun to explore centrality measures<\/em> with relatable data (hint Valentines Day)\u00a0to demystify meaningful experiences common to social networks.\u00a0 \u00a0How does Pandora music know that I love The Strokes<\/em>?\u00a0 \u00a0Even though it seems like the station owner may have lived within earshot of my speakers, these personalized suggestions are actually a result of a graph analysis with centrality measures that factor in an entire platform of users.\u00a0 These measures then help tailor different software experiences.\u00a0 Within Insights, you can build graphs and calculate centrality measures to deliver meaningful recommendations based on relationships within your own data.<\/p>\n

This is powerful.\u00a0 It’s powerful because decisions that involve link analysis<\/em> typically factor in complex relationships which are difficult to see in classic row-and-column data structures.<\/p>\n

Even though many of us will use Insights to explore linkages from different industry domains, the aim in this blog post is to keep things simple and apply research in an arena where link analysis<\/em> is deeply rooted.\u00a0 No, it’s not epidemiology or forensics (even though it could be).<\/p>\n"},{"acf_fc_layout":"image","image":{"ID":439032,"id":439032,"title":"vday-fig3","filename":"vday-fig3.png","filesize":229250,"url":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2019\/02\/vday-fig3.png","link":"https:\/\/www.esri.com\/arcgis-blog\/vday-fig3","alt":"","author":"8962","description":"","caption":"","name":"vday-fig3","status":"inherit","uploaded_to":0,"date":"2019-02-14 17:31:12","modified":"2019-02-14 17:31:12","menu_order":0,"mime_type":"image\/png","type":"image","subtype":"png","icon":"https:\/\/www.esri.com\/arcgis-blog\/wp-includes\/images\/media\/default.png","width":2714,"height":1723,"sizes":{"thumbnail":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2019\/02\/vday-fig3-150x150.png","thumbnail-width":150,"thumbnail-height":150,"medium":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2019\/02\/vday-fig3-300x190.png","medium-width":300,"medium-height":190,"medium_large":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2019\/02\/vday-fig3-768x488.png","medium_large-width":768,"medium_large-height":488,"large":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2019\/02\/vday-fig3-1024x650.png","large-width":1024,"large-height":650,"1536x1536":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2019\/02\/vday-fig3.png","1536x1536-width":1536,"1536x1536-height":975,"2048x2048":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2019\/02\/vday-fig3.png","2048x2048-width":2048,"2048x2048-height":1300,"card_image":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2019\/02\/vday-fig3-732x465.png","card_image-width":732,"card_image-height":465,"wide_image":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2019\/02\/vday-fig3-1701x1080.png","wide_image-width":1701,"wide_image-height":1080}},"image_position":"center","orientation":"horizontal","hyperlink":""},{"acf_fc_layout":"content","content":"

Given that it’s Valentines Day, we’ve decided to see how Link Charts and Pre-Defined Filters can be applied in a similar fashion to popular dating app features – aka, to find love interests (but without the swipes).\u00a0 We will use Insights to sift through online dating data and see how well ArcGIS can find matches for companionship.<\/p>\n

Keep in mind our key message here is not about how to be a busybody.\u00a0 \u00a0Rather the take away is how Link Charts can slice apart descriptive information and reveal hard-to-see visual linkages and centrality degrees, which make data actionable.\u00a0 In Insights we do this by enabling different ways to style nodes and links in charts using color, sizes, line thickness and layouts (force directed, circular and hierarchical).\u00a0 For the rest of this blog we will romantically analyze Bill.\u00a0 Bill is a fictitious person (located in the center of the link chart<\/em> above) who is overwhelmed by a large online dating pool of women who have different tastes, desires and preferences.\u00a0 Does this sound familiar?\u00a0 \u00a0Perhaps on several levels, it should.\u00a0 This is a similar problem that practitioners will face when it’s time to make a needle-in-a-haystack type recommendation and they are overrun with information overload.<\/p>\n

Data Gathering Process<\/h2>\n

For our research we have analyzed open data collected by a popular dating site named OkCupid using Insights.\u00a0 \u00a0The nature of the data includes 68,000’ish anonymous responses to some 2,000+ self-assessment questions.\u00a0 \u00a0Per OkCupid, this sampling is the exact type of data they use to build their matching algorithms.\u00a0 While we did not have time to review the\u00a0 interesting R scripts and reports packaged with the data, on the whole we will attest that OkCupid’s sample data collection is pretty awesome and the Q&A topics were pretty fun to read.\u00a0 Here is a direct link to the data<\/a>.<\/p>\n

 <\/p>\n"},{"acf_fc_layout":"image","image":{"ID":439172,"id":439172,"title":"vday-fig2-final-final-final","filename":"vday-fig2-final-final-final.png","filesize":140904,"url":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2019\/02\/vday-fig2-final-final-final.png","link":"https:\/\/www.esri.com\/arcgis-blog\/vday-fig2-final-final-final","alt":"","author":"8962","description":"","caption":"","name":"vday-fig2-final-final-final","status":"inherit","uploaded_to":0,"date":"2019-02-14 18:20:57","modified":"2019-02-14 18:20:57","menu_order":0,"mime_type":"image\/png","type":"image","subtype":"png","icon":"https:\/\/www.esri.com\/arcgis-blog\/wp-includes\/images\/media\/default.png","width":1906,"height":1569,"sizes":{"thumbnail":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2019\/02\/vday-fig2-final-final-final-150x150.png","thumbnail-width":150,"thumbnail-height":150,"medium":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2019\/02\/vday-fig2-final-final-final-300x247.png","medium-width":300,"medium-height":247,"medium_large":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2019\/02\/vday-fig2-final-final-final-768x632.png","medium_large-width":768,"medium_large-height":632,"large":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2019\/02\/vday-fig2-final-final-final-1024x843.png","large-width":1024,"large-height":843,"1536x1536":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2019\/02\/vday-fig2-final-final-final.png","1536x1536-width":1536,"1536x1536-height":1264,"2048x2048":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2019\/02\/vday-fig2-final-final-final.png","2048x2048-width":1906,"2048x2048-height":1569,"card_image":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2019\/02\/vday-fig2-final-final-final-565x465.png","card_image-width":565,"card_image-height":465,"wide_image":"https:\/\/www.esri.com\/arcgis-blog\/app\/uploads\/2019\/02\/vday-fig2-final-final-final-1312x1080.png","wide_image-width":1312,"wide_image-height":1080}},"image_position":"center","orientation":"horizontal","hyperlink":""},{"acf_fc_layout":"content","content":"

Data adjustments and alterations<\/h3>\n

To save some of your time, I’ve outlined our data munging process.\u00a0 Since the OkCupid data is completely anonymous, we made a few changes to add a bit of context.<\/p>\n