4. … or the end station is full and
can't find empty docks!
5. The Problem: Planning a ride
-Rush hour:
Probability(station has bikes): 100% to 20% in 2hr!
User needs:
-Best start stations with an available bike?
-Best end stations with an available dock?
7. Algorithm:
-Build indexed MySQL database with 1 year data.
(30GB)
-Fetch similar trips on the fly:
-within [T - 15min, T + 15min]
-repeat within [T – 15days, T + 15days]
-separate weekdays from weekends/holidays.
8. Algorithm:
-Average of bike/dock availability.
-Quantiles of travel times.
-Ranking stations:
SCOREi=
Pi(stationhas bikes)
C1+C2∗WalkDistancei
9. Manuchehr Taghizadeh-Popp
-Astrophysics PhD
Johns Hopkins University
~10TB Galaxy databases
-Galaxy classification
-Eigen-images
-Stats of Extreme Values
-Universe simulations