Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Big Data Envy: A Chief Data Officer's Journey

Avatar for Greg Elin Greg Elin
February 27, 2013

Big Data Envy: A Chief Data Officer's Journey

Not all agencies have exabytes of data. But size does not matter. Big Data is really about having the right data fast enough to meaningful model complex systems for cause and effect. See how one Chief Data Officer looks at the challenge.

Avatar for Greg Elin

Greg Elin

February 27, 2013

More Decks by Greg Elin

Other Decks in Technology

Transcript

  1. CHIEF DATA OFFICER, FCC (2010 - 2013) Creator of Sunlight

    Foundation’s Sunlight Labs (2006 - 2009) Greg Elin @gregelin Big Data Envy! A Chief Data Officer’s Journey
  2. Big Data Envy Small Data Life Same Data Pain Data

    deluge Data for record keeping Analysis ready? Real time Periodic Insight for decision Too much data Too much variation Too costly to move Loosely structured Loosely structured Costly to standardize Heterogeneity Heterogeneity Scaling challenges Distributed across networks Distributed across networks Permission nightmare Complex interrelationships Complex interrelationships Constantly redefined Diverse sources Diverse sources Time to organize “Big Data = Difficult to capture, store, search, share and analyze” NIST
  3. Defining Big Data Data sufficient in scope, detail and timeliness

    to model dynamic system for cause and effect. See. Predict. Choose. Influence.
  4. Boyd’s OODA Loop Deming Cycle PDCA Motorcycle Safety SIP-DE Observe

    Orient Decide Act ! Plan Do Check Act (Adjust) Scan Identify Predict Decide Execute Big Data supports situational awareness decision making. See. Predict. Choose. Influence.
  5. Modernize and streamline how the FCC: Collects Data EFFICIENT, TIMELY,

    RELEVANT Uses and Manages Data STANDARDIZE AND MODERNIZE Shares Data TRANSPARENT, USEFUL PUBLIC DATA TOOLS
  6. Roadmap  to  Stats  (APIs  /  Dashboards) 1.Show  examples  and  discuss

      2.Iden2fy  just  6  to  12  ques2ons   3.Gather  the  SQL  statements  into  a  doc   4.Pick  friendly  URL  to  display  answers   5.Repeat  and  extend