Big Graph Data Intelligence – Analyzing Large Connected Data in Social and Health Science

September 18, 2018 | 6pm to 8.30pm | WeWork 71 Robinson

Abstract

The miracle behind Social Data is that we have information on the detailed structure of how people are connected to each other, who are the family members, who are the friends, and who are the influencers, who are the followers. We call this structure as graph topology. But beside the Social Data Topology (from diverse sources) we can also observe characteristics of and behavior mechanism among the individuals. Many researches have proved that inside micro-topologies (cliques or communities) people tend to think, decide, purchase or do similar things, have similar profiles in many cases. Thus, if we want to understand or change customers’ decision, we must use the micro-topology information, not just individual connections.

Today, multinational companies (banks, airlines, telecoms, insurance companies and many other domains) are closer than ever to analyze, understand and utilize Social Data. They usually have at least three different sources of Social Data, which is sufficient to build their own transactional Social Network. First, A-to-B transactions (calls, instant messages, money transfers, bookings). Second, (co-)locations and the (co-)movements (same address, sharing bills, traveling together). Third, digital behavior (browsing history, app usage) of customers, potentially complemented with external information. Just like Facebook helps not only to derive the Network, but add much diverse information on attributes and interests. This in turn enables further deep dive on the homogeneity of these communities, verify those cases when the network is similar in demography, when the network is giving an insight on commercial decisions, when the network enables the spread of word of mouth.

However, creating the transactional Network using these attributes is not easy, especially when there are dozens of millions of customers, and billions of possible and measurable interactions of and between them. The different sources of information can contain contradictions and confusions, when we observe that overlaid topologies are not matching. (E.g.: online friends and offline friends.) Lynx Analytics has experiences and solutions to build and use these networks efficiently.

Speaker

Gábor Benedek is an innovation partner at Lynx Analytics providing predictive analytics for communication companies, financial services and healthcare sectors in South-East-Asia. He has been applying SNA methodologies for Celcom, Indosat, Singtel, Telkomsel, Globe, DBS Bank in the region. Gábor received his PhD in 2003, in 2012 his T-Mobile SNA churn study was awarded as the Best Application paper by the European Decision Science Institute, and he is the author of one book and author/coauthor of over 20 articles. He was an Associate Professor at Corvinus University of Budapest, researching and lecturing in the fields of economic and business simulations, social network analysis, econometrics, data mining and predictive analytics. Gábor was among the founders of Data Explorer, the first consulting company in predictive analytics in Hungary. In 1999 Data Explorer built the first social network analysis software for churn and community detection applicable for mobile customers in Europe, based on Gábor’s theoretical foundations and proposals. In 2010 Gabor was contributing to the largest public physicians’ social network in the world, based on real patient-flow data between general practitioners and specialists.

Date/Time Tuesday 18 September 2018, beginning with a light reception at 6pm

Venue WeWork 71 Robinson Road Singapore 068895

RSVP on our Meetup.

See you there!