Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

MeetMeWhere

126 views

Published on

Insight data science presentation.

Published in: Science
  • Login to see the comments

  • Be the first to like this

MeetMeWhere

  1. 1. Meet You Where? Josh Friedman
  2. 2. Meet You Where? Josh Friedman
  3. 3. Meet You Where? Josh Friedman
  4. 4. Meet You Where? Josh Friedman
  5. 5. Meet You Where?
  6. 6. Meet You Where? ?
  7. 7. Meet You There 1.Identify an equitable meeting location. I Travel 10 minutes They Travel 10 minutes
  8. 8. Meet You There 1.Identify an equitable meeting location. I Travel 10 minutes They Travel 10 minutes
  9. 9. Meet You There 1.Identify an equitable meeting location. I Travel 10 minutes They Travel 10 minutes
  10. 10. Meet You There 1.Identify an equitable meeting location. 2. Finds restaurants near the ideal meeting place.
  11. 11. Meet You There 1.Identify an equitable meeting location. 2. Finds restaurants near the ideal meeting place. 3.Identifies the best ones for you. • Travel fairness ratio. • Best reviewed. • Lowest Price.
  12. 12. Behind the scenes Traffic Data Calculate Midpoint
  13. 13. Behind the scenes Search Location Traffic Data Calculate Midpoint
  14. 14. Behind the scenes Ranked Locations 1. Ray's Pizza 2. World-Famous Original Ray's Pizza 3. Ray's Original Pizza 4. Famous Ray's Pizza Search Location Cosine Similarity Traffic Data Calculate Midpoint
  15. 15. Behind the scenes Ranked Locations 1. Ray's Pizza 2. World-Famous Original Ray's Pizza 3. Ray's Original Pizza 4. Famous Ray's Pizza User Selection: Search Location User, Review, Price, Travel Fairness Ratio Cosine Similarity Traffic Data Calculate Midpoint
  16. 16. Behind the scenes Ranked Locations 1. Ray's Pizza 2. World-Famous Original Ray's Pizza 3. Ray's Original Pizza 4. Famous Ray's Pizza User Selection: Search Location User, Review, Price, Travel Fairness Ratio Cosine Similarity Traffic Data XCalculate Midpoint
  17. 17. The Cold Start Problem – Where do people tend to go? More Expensive Average Price (x) and Review (y) selections for 30K people. Higher Rating
  18. 18. Spur of Moment The Cold Start Problem – Where do people tend to go? Gourmets More Expensive Higher Rating + + +
  19. 19. MeetMeWhere.xyz
  20. 20. The Cold Start Problem – Where do people tend to go?
  21. 21. Rank Locations by Cos Similarity to User Profile 𝜈local = (R・vi ) + (P・vj)+ (T・vk) 𝜈user = ( 𝛚R・vi ) + ( 𝛚 P・vj)+ ( 𝛚 T・vk) Place Review (1-5) Travel Fairness (0-100%)Price Range (0-4) Rank = 𝜈local ・ 𝜈user
  22. 22. Meet You Where? I travel 10 min They travel 10 min 1. Calculate the mid-point based upon projected real time traffic data or public transit delays (google)
  23. 23. Meet You Where? I travel 10 min They travel 10 min 1. Calculate the mid-point based upon projected real time traffic data or public transit delays (google)
  24. 24. Meet You Where? I travel 10 min They travel 10 min 1. Calculate the mid-point based upon projected real time traffic data or public transit delays (google)
  25. 25. Meet You Where? I travel 10 min They travel 10 min 1. Calculate the mid-point based upon projected real time traffic data or public transit delays (google)
  26. 26. Meet You Where? 1. Search the vicinity of the mid-point for establishments matching the user query (google places API)
  27. 27. Meet You Where? The Best One! … But what's important? Review? Price? Distance? Some combination?
  28. 28. Behind the scenes Ranked Locations 1. Ray's Pizza 2. World-Famous Original Ray's Pizza 3. Ray's Original Pizza 4. Famous Ray's Pizza User Selection Search Location User, Review, Price, Travel Fairness Ratio Cosine Similarity User Preferences Traffic Data Calculate Midpoint

×