Data Mashups -Data Science Summit

3,522 views
3,417 views

Published on

As large datasets come together exciting and unexpected things can happen. Human behavior is high dimensional, so combining many diverse datasets is critical to revealing actionable insights.

Published in: Technology
0 Comments
5 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
3,522
On SlideShare
0
From Embeds
0
Number of Embeds
149
Actions
Shares
0
Downloads
43
Comments
0
Likes
5
Embeds 0
No embeds

No notes for slide

Data Mashups -Data Science Summit

  1. 1. Data Mashups May 12, 2011 Data Scientist Summit Turning Data Exhaust into Pete Skomoroch LinkedIn Insights @peteskomoroch
  2. 2. We have an explosion of data • DataWrangling • InfoChimps • Data.gov • Factual • SimpleGeo
  3. 3. And the tools to make sense of it • Hadoop • NoSQL •R • Python • Mechanical Turk
  4. 4. Diverse datasets = better signal
  5. 5. Find a meaningful problem • Identify pain points • Work on stuff that matters • Focus on underutilized data http://www.flickr.com/photos/aloshbennett/
  6. 6. Trendingtopics.org @hourlytrends
  7. 7. LinkedIn Skills
  8. 8. The best mashups are actionable • Reveal patterns • Enable predictions • Recommendations
  9. 9. Mashup: Skills & Cities
  10. 10. Yuba City, California: 21.3% Unemployment
  11. 11. Ames, Iowa: 4.7% Unemployment
  12. 12. Make data mashups work for you • Open Data = powerful mashups • Mashup > sum of its parts • Focus on meaningful problems • Actionable mashups are better

×