Takeaway Tales

65bea007fa26257adff7aaf5b7268e09?s=47 peteowlett
February 09, 2016

Takeaway Tales

Talk given at @datamates London about the weird things we can learn about the world by looking at Deliveroo's takeaway data.

65bea007fa26257adff7aaf5b7268e09?s=128

peteowlett

February 09, 2016
Tweet

Transcript

  1. TAKEAWAY TALES

  2. Hello! I’m Pete

  3. None
  4. I work for these folks

  5. WE DO THIS

  6. Chapter 1 Your new year’s resolution is radioactive

  7. People cut down on takeout at Christmas

  8. None
  9. Something i learned in school

  10. Looks familiar ….

  11. Strength … failing …

  12. Yep, highly predictable

  13. The Half Life of New Years Resolutions is 11 Days

  14. Chapter 2 Why Notting Hill is like a Class G

    Star
  15. CLASS G

  16. FANCY EH?

  17. Turns out, Delivery Zones are like that too Througput (orders

    per driver per hour) # LOG Orders over 91 days
  18. Turns out, Delivery Zones are like that too Througput (orders

    per driver per hour) # LOG Orders over 91 days Notting Hill
  19. None
  20. Chapter 3 You can learn a lot from people’s data

    exhaust
  21. We were looking at customer buying patterns …

  22. Lets look at our most active customers

  23. We learn from events, but we ALSO learn from the

    absence of events
  24. Chapter 4 Is restaurant variety the spice of life?

  25. Some People Order From One Restaurant Repeatedly

  26. These people are more likely to leave

  27. Lets Dig Deeper

  28. The data says we’re wrong - restaurant variety doesn’t decrease

    churn
  29. WHAT HAVE WE LEARNED?

  30. AT SCALE, PEOPLE ARE PREDICTABLE

  31. THINK ABOUT WHAT YOU SHARE

  32. ACCOUNT FOR BIAS

  33. THANKS! @PeterOwlett