Chicago Taxi Trips (Tipping)

Published: 22-06-2021| Version 3 | DOI: 10.17632/tm47d7z5xf.3
Sarah Conlisk


This dataset accompanies a paper examining how tipping practices changed during the pandemic. I download data on unique taxi rides taken in Chicago freely available from the Chicago data portal for Jan 2018-March 2021. The data include a number of variables for each taxi trip, including the fare amount, the tip amount, and the passenger's location. I merge in demographic data available from CMAP for the community area the passenger came from in Chicago as well as daily data on COVID-19 hospitalizations to enrich the analysis. I filter the dataset to taxi rides payed with credit card, and remove trips with exceptionally weird data (e.g. 0 second trip duration, fares greater than $1000). I use the dataset to estimate the effect of the pandemic on whether passengers tip and if so, the average percent tipped. I find that the likelihood that a passengers tips during the pandemic is about 5 percentage points lower but the percentage tipped of the fare when they do is about 2 percentage points higher.


Steps to reproduce

1. Download taxi data available from the Chicago data portal for years 2018, 2019, 2020, and 2021 2. Download the 2 Chicago community areas datasets and the COVID-19 Hospitalizations data available here 3. Run file Taxi_clean.R to clean the data and generate calendar heat maps. This code will produce the final dataset "Taxi_Trips.csv" also available from here. 4. Run file to repeat the regressions and charts.