Help, I’m the first data scientist at my company!

I moderated a panel at DataGotham with Adam Laiacano from TumblrFred Benenson from Kickstarter, and Roberto Medri from Etsy about being the first data scientist at a company. We covered everything from what people’s job responsibilities are, the tools they use, successes, failures, how they are integrated into an organization, and how they have hired other data scientists to join them. The panelists were concise, articulate, and intelligent. Watch it below!

  • Pingback: Panel Discussion of Data Scientists | Anton Schieffer

  • Steve Akers

    About halfway through the discussion the comment was made that StatsD and Graphite are only useful for looking at whether stuff is going up or down. That’s not entirely correct. It’s good at that, but can be used for much more. At LivingSocial we’re building a real-time monitoring framework that allows us to analyze Graphite data beyond basic trending. For example, we run ANOVA before and after a deployment to determine its impact on metrics like response times and transaction counts (e.g. logins, purchases, emails sent, etc.) We also generate control charts comparing current data to same time/day previous week to determine when a process has exceeded acceptable control limits. All this is done in near real-time. As we increase our monitor coverage it will free us up to make more frequent changes to our systems.