Discover relationship patterns: communities, clusters, central nodes  This talk is for Dojo members only. If you want to join this event, please register at the Dojo Front Desk or send us an email to events@dojobali.org. Places are limited. Introduction: Network graphs, such as the connections between users on Facebook or LinkedIn, are not well suited to traditional relational databases, like SQL. In relational databases, each row represents a “thing,” such as an event or item. The emphasis is generally on finding, storing, and analyzing relationships between these items or events. More technically, in a relational database, an item or event is a first order citizen.  In graph databases, a relationship is the first order citizen. Each row, for example, represents a relationship, as opposed to an item. This enables analysis based on these relationships, such as identifying communities, clusters, and strengths of relationships. In this talk, we’ll use an example/analysis based on the MedLit database, which provides an aggregation of publications and bibliographies across the BioTech industry. This data was then analyzed and converted to a graph database, representing relationships between researchers (e.g., who published or worked with whom). The subsequent analysis will demonstrate some of the kinds or relationship patterns that can be identified using Network Graph Analysis tools and techniques, such as identifying communities and clusters of researchers that tend to work together, identifying central nodes that tend to either bridge, or identifying nodes with high power/centrality that tend to interact with a large number of other nodes. Eventually, this provides the tools and laguage to describe and compare networks. For instance, would this network of researchers look similar to a LinkedIn network of sales managers? Or, for example, do two people have similar types of networks of friends on Facebook; and, if not, what types of network patterns represent what types of real world relationships? What you will learn: – how to find, store, and analyze relationships and their patterns- how to identify communities, clusters and central nodes- how to compare networks About Denis: Denis is a Data Scientist and Project Manager. Most recently, he worked at Sandisk on marketing and in-app analytics big data projects, developing an in house BI tool to enable product marketing managers to explore and analyze terabyte size datasets. He is currently teaching at UC Berkeley, helping with the Python/Intro to Data Science summer courses, while also developing new content, including Agile Scrum for Data Scientist and an Intro to Network Graph analysis seminar. Prior to moving to data science, he worked as an entrepreneur in the International Affairs domain. He has a background in Engineering Physics and International Relations. Looking forward to seeing you there!