Traffic Jam: Predicting People's Movement into Nairobi
$12,000 USD
Uber and Mobiticket team up to predict demand for public transportation into Nairobi
6 September 2018–13 January 2019 23:59
690 data scientists enrolled, 204 on the leaderboard
Can not replicate results from Uber data
published 6 Nov 2018, 08:08
edited less than a minute later

Hi, Uber gives you the option to download "Filtered Data" for your specific date range, for a specific route. The files you can download are:

1. travel_times_daily.csv

2. travel_times_day_of_week.csv

3. travel_times_time_of_day.csv

So I downloaded these files for April-June 2017 for a specific route. If I am correct, files 2 and 3 can be generated using data from file 1. I did that, but strangely, my versions of files 2 and 3 do not match the values of uber's 2 and 3. For example, in travel_times_daily.csv, there is no data for the column "PM Mean Travel Time (Seconds)". Yet, travel_times_time_of_day.csv shows non-zero values for "PM Peak". How on earth is that possible? There are many other (and more serious) inconsistencies I found with travel_times_day_of_week.csv file. I think Uber data is misleading. Is it possible to contact them?