Distributed Data Show Episode 80: Finding Bad Actors with Max Melnick
Highlights
0:15 -You told us you want to hear about use cases, so here we go!
0:58 - Getting to know Max - Virginia native, TensorFlow committer, Analytics at Deloitte
2:35 - The problem space for Mission Graph - identifying high-risk actors in networks aka "bad guys" to help analysts be more effective
3:53 - Mission Graph - solving not only the data problem of high volume, but the operational problems of connecting/correlating across data sets
5:35 - Mission Graph is a self-service platform for interactive network exploration and analysis. With current systems, most analyst time is spent on data engineering tasks. Mission Graph is trying to fix this.
7:58 - The technology stack includes Cassandra, DataStax Search, DataStax Graph. Lessons learned include: 1) avoiding network-based storage and doing regular health checks.
9:48 - 2) When onboarding new team members, give them initial tasks that correspond to their skill level.
11:17 - 3) When using Graph, use the DSE GraphFrames API for large volume data loading. Use Spark GraphFrames to take advantage of algorithms page rank, motif finding, etc.
13:28 - Wrapping up