Datathon - Big Data Hack-a-Thon - Workshop

This is a past event

92 people went

CandiDate / Industrious Office

230 S. Broad Street - 17th Floor · Philadelphia, PA

How to find us

Entrance is on Locust Street, between Broad and 15th

Location image of event venue

Details

Datathon was created by DataPhilly (http://www.meetup.com/dataphilly) with partnership with Code for Philly (http://www.meetup.com/code-for-philly/) as a voluntary effort for the technology community to come together and combine software development, machine learning and predictive analysis expertise to develop a service that helps researchers identify drugs that are targeted to the genomic features of specific cancers. Read more (http://www.meetup.com/DataPhilly/events/230814092/).

The Workshop

You will need your computer for this event.

This workshop session will give the experienced designers, engineers, and data enthusiasts participants an opportunity to share their expertise in technologies and/or best practices that they have employed in their professional experience or their own projects.

​*Speaker bio*​: Daniel Himmelstein is a Postdoctoral Fellow at Penn. He develops hetnets — networks with multiple node and relationship types — to encode decades of biomedical study into a single resource. His latest hetnet, named Hetionet, contains 2,250,197 relationships of 24 types and is being applied to predict new uses for existing drugs.

​*Topic*​: We will introduce a graph database, Neo4j, as well as the graph query language Cypher. We'll discuss how to represent complex data as hetnets in Neo4j. We'll also introduce Hetionet (http://neo4j.het.io/) — a network of biomedical knowledge containing 11 node types including drugs, diseases, genes, side effects, and symptoms. We'll go over a machine learning approach to repurpose drugs that is designed for hetnets (referred to as Project Rephetio (https://thinklab.com/discussion/predictions-of-whether-a-compound-treats-a-disease/203)).

​*Skill-level*​: Basic programming and data science skills are recommended. We won't assume any Neo4j or biology experience. The workshop is recommended for those interested in data integration, networks, biology, medicine, data science, or open data.

Schedule

6pm - Doors open

6:45 - Civic Hacking 101

7pm - Lightning talks and workshop session

9pm - Commit and quit

You will need your computer for this event.