Title: Decoding the Urban DNA and Harnessing the Power of Social Intelligence
1Decoding the Urban DNA and Harnessing the Power
of Social Intelligence
May 2009 Where2.0 Greg Skibiski,
CEO greg_at_sensenetworks.com
2 Sandy Pentland Chief Privacy Advocate MIT,
Head of Human Dynamics Research
Tony Jebara Chief Scientist Columbia
University, Head of the Machine Learning
Program
3Static Demographics
4Static Demographics
5HOME
6(No Transcript)
7Half the people on the planet are creating
location data It is the largest, most universal
source of information describing human behavior
8Goal understand people
- Its hard to say if User A is like User B from
location data - People dont overlap in space time enough to
compare, but they do overlap semantically once we
understand places
User A
User B
9First, understand places
Is place A like place B? Look at each places
Flow, Commerce Demographics
10(No Transcript)
11noon saturday
8pm saturday
8pm tuesday
noon tuesday
12(No Transcript)
13Dynamic Demographics
- Matrix of 9 example users - original location
data can be deleted -
14Macrosense 1.0 Prediction Platform
Learn
Predict
Prediction algorithms trained on a sample of
users (n10,000) Model generated automatically
and performance evaluated on unseen data
Models are deployed at scale (n10,000,000)
Latitude, Longitude, Time, Binary Business
Prediction
Predictive Model
Latitude, Longitude, Time
Binary Business Variables (x,y)
15Dynamic Demographics
-
- Recommendation
- Personalization
- Discovery
- 2x Improvement
16Dynamic Demographics
17Dynamic Demographics
Local Search
18Privacy Data Ownership
- All opt-in
- Location data is not retained after processing
- No traces kept
- k-level anonymized
- New Deal on Data thought leadership
- You have the right to possess your data
- You have full control over the use of your data
- You have a right to dispose or distribute your
data
19(No Transcript)
20Decoding the Urban DNA and Harnessing the Power
of Social Intelligence
May 2009 Where2.0 Greg Skibiski,
CEO greg_at_sensenetworks.com