NYC traffic

• Goal: reduce accidents
• Intervention: eg new street signs
  • To what extent will these work and how much do you spend?
• Could be more effective
  • e.g. Broadway has a lot of accidents (tho not fatalities)
  • other locations have fewer accidents, more fatalities
  • where should cops be???
• what are the characteristics of intersections where accidents happen?
• Goal: having Uber reduces car uses
  • Observation: Uber = more cars and less transit
• Data availability
  • avaiable but still difficult to get and analyze

BC Transit

• Target areas where people come and go - high traffic
  • count people! indicates greater and lesser use
  • cards / phone app
• Modeling high traffic areas
  • Can we predict areas that need a lot of transit?
  • Take high transit locations and lot of info
  • Build model - what’s predictive
  • Make sure the model works
• Twitter data
  • Improve locations tagged as not performing well
  • Keep in mind that not everyone tweets
    • younger
    • likely more well off