Chicago Taxi Trips (Tipping)
Description
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 passenger leaves a tip declines by roughly 5 percentage points during the pandemic but the average non-zero tip left increased by roughly 2 percentage points higher. I exploit geographic and temporal heterogeneity in the data to explore the possible mechanisms behind these patterns in tipping.
Files
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 Taxi_tip_analysis.do to repeat the regressions and charts.