Nashville Scooter Usage - Multiple

Team Members: Roy Lugos, Diarra Fall, Aidan Dubois, and William Prunty

The Nashville SUMD's project was an investigation into usage patterns of shared urban mobility devices, specifically the scooter sharing program.

The goal was to find what density of scooters would best service the area while limiting the pile up of unused scooters on public sidewalks. 

The Nashville SUMD's project utilizes several technologies with the majority of the technical analysis being performed in Python. Using Pandas, GeoPandas, Folium, SQLalchemy,  matplotlib, and others to perform the analysis on a combined ~100 million datapoints. Additionally, data was brought into Tableau and Excel for visualizations. 

heatmap2x.mp4

Heatmap created in Tableau showing the trip originations for a given week by hour, allowing the viewer the opportunity to see the flow of scooter traffic through time.

Series of Python created visuals analyzing scooter distance and duration distribution. Click through for the full format images.

Findings - Scooter Usage

The primary goal of this project was to make recommendations on the proper density of scooters to service trips and reduce the number of unused scooters on public sidewalks. After thoroughly investigating the dataset, we drew the following conclusion:

"Scooter density for the zip codes representing the highest scooter utilization,  reaches a peak density of  ~440 scooters per square mile. Density is roughly half that for weekdays.

With the current density, only 1/3 of scooters are reaching 'sufficient use' as defined by servicing  3 or more trips per day, another 1/3 are 'underused' or servicing at least 1 trip per day. The final 1/3 go unused.

The trend is identical in the more 'rural' areas, however fleet size is much smaller.

Generally, we recommend reducing fleet size by 1/3 in the downtown area, which are also the areas servicing the majority of SUMD trips. This allows for some growth of the scooter program while keeping  un- or under- utilized scooters from cluttering the street. If the interest is in increasing utilization of scooters as a means of transportation outside of the downtown area, we recommend not altering the scooter fleet size in those areas.

A dynamic density could be considered, where more scooters are made available during the weekend. Our recommendation for a dynamic model would be to have ~300 scooters available during the weekday and ~500 available during the weekends and again maintaining the same fleet size outside of these areas to encourage adoption of shared mobility devices. "

GitHubLinkedIn