inferring travel from social media
play

Inferring Travel from Social Media Alessio Signorini - PowerPoint PPT Presentation

Inferring Travel from Social Media Alessio Signorini <alessio-signorini@uiowa.edu> Alberto Maria Segre <alberto-segre@uiowa.edu> Philip Polgreen <philip-polgreen@uiowa.edu> ONCE UPON A TIME... H1N1 TWEET VOLUME CDC


  1. Inferring Travel from Social Media Alessio Signorini <alessio-signorini@uiowa.edu> Alberto Maria Segre <alberto-segre@uiowa.edu> Philip Polgreen <philip-polgreen@uiowa.edu>

  2. ONCE UPON A TIME...

  3. H1N1 TWEET VOLUME CDC recommends canceling travels plans Pandemic level raised to 5 Number of confirmed cases reach 1000

  4. AMERICAN IDOL 2009 More Positive Tweets about Kris Allen 45 40 ALLEN 35 % positive tweets 30 vs. 25 20 15 10 5 LAMBERT 0 days

  5. REAL-TIME ILI% ESTIMATE Reported vs. Predicted Weekly ILI% Flu Season 2009-2010 - United States 8.5 8 7.5 7 6.5 6 5.5 5 4.5 % ILI 4 Predicted 3.5 3 2.5 2 1.5 Reported 1 0.5 0 09/40 09/41 09/42 09/43 09/44 09/45 09/46 09/47 09/48 09/49 09/50 09/51 09/52 10/01 10/02 10/03 10/04 10/05 10/06 10/07 10/08 10/09 10/10 10/11 10/12 10/13 10/14 10/15 10/16 10/17 10/18 10/19 10/20 1-fold validation ~ error avg=0.28%, min=0.04%, max=0.93%. Std=0.23%

  6. DEFINITELY A GOOD IDEA!

  7. SICK PEOPLE STILL TRAVEL + + CURRENT FLU MAP TRAVEL MODEL

  8. TRAVEL MODELS CENSUS TRAFFIC TICKETS MONEY CELL PHONES

  9. TRAVEL MODELS CENSUS TRAFFIC TICKETS MONEY CELL PHONES

  10. TRAVEL MODELS CENSUS TRAFFIC TICKETS MONEY CELL PHONES

  11. TRAVEL MODELS CENSUS TRAFFIC TICKETS MONEY CELL PHONES

  12. TRAVEL MODELS CENSUS TRAFFIC TICKETS MONEY CELL PHONES

  13. TRAVEL MODELS CENSUS TRAFFIC TICKETS MONEY CELL PHONES

  14. GPS ADDED TO CELL PHONES STEVE JOBS

  15. GPS ADDED TO CELL PHONES STEVE JOBS

  16. LOCATION-BASED APPS

  17. FOCUSING ON THE MOST POPULAR CHECK-IN TO PLACES TO EARN BADGES TWEETS CAN BE GEO-LOCATED

  18. FOLLOW PEOPLE EVERYWHERE RESTAURANT BAR DOCTOR GYM OFFICE STARBUCKS

  19. DATA COLLECTED Number of Locations 76 MILLION Number of Users 6 MILLION

  20. DATA CLEANUP CASUAL USERS TOO FREQUENT TOO FAST (<5 locations) (>1 every 5 secs) (>1800 km/h)

  21. DISTANCE TRAVELLED 100% 99% 97% 100% 85% 90% 80% 70% 60% 50% 50% 40% 30% 20% 10% 0% 0 < 1 mile 1 < 10 miles 10 < 100 miles 100 < 1000 miles 1000 < 10000 miles % Trips % Cumulative

  22. TIME INTERVAL 97% 16% 89% 14% 81% 12% 69% 59% 10% 46% 8% 38% 6% 31% 24% 21% 4% 15% 8% 2% 4% 1% 0% 0% 10s 30s 1m 2m 5m 10m 15m 30m 1h 2h 6h 12h 1d 2d 1w % Trips % Cumulative

  23. TRIPS vs. DISTANCE 22.3 22.0 21.6 21.2 20.8 19.8 17% 18.5 16% 16% 14% 14% 12% 11% Monday Tuesday Wednesday Thursday Friday Saturday Sunday % Trips Miles

  24. TYPICAL NEW YORK CITY DAY 6 AM 2 PM 8 PM

  25. TRACKING INDIVIDUALS

  26. AGGREGATES BETWEEN U.S. STATES

  27. WHERE TO GET MORE INFORMATION Alessio Signorini alessio-signorini@uiowa.edu http://www.cs.uiowa.edu/~asignori/ UIOWA Computational Epidemiology Group http://compepi.cs.uiowa.edu paper and datasets will be soon available on the CompEpi website

Recommend


More recommend