Bike Sharing Demand in San Francisco
20 July, 2020
3
3
0
Contributors
Overview
Studying the San Francisco bike-share dataset on Google BigQuery, we can see a pattern in usage and how this can affect the availability of bikes in various stations. Looking over the dataset, we can find that there a total of 13 stations that do not have any bikes available. However, 10 of the 13 are not installed on the street yet. Out of the three remaining stations, station ID 146 is currently not renting, which would make station IDs 295 and 212 as the only location with no bikes available and is currently renting. 19 stations have a cumulative total of 76 bikes available, but the station is currently “not renting.” Since these stations are not renting, the bikes are sitting idle with no usage.
Breaking down capacity and bike availability, we see that averaging all stations, 56% of bikes are not being used based on their station capacity - more than half of the bikes are not used. Drilling the details down further, there 218 stations that have a higher capacity of available bikes than the average 56%, which equates to 3614 not used bikes. Within this group, there are 65 stations with over 80% available bike capacity – that averages 21 available bikes per station. These stations are not being used as much and could have their bikes moved elsewhere.
Shifting focus on how demand changes based on locations. If we look at trips that had the same pickup and drop-off location, the highest demand is around the Embarcadero area in downtown San Francisco. 4 out of the top 6 stations have over 1000 rentals, with 3 stations accounting for closer to 2000 trips each. A continuing focus of the 4 main stations in the Embarcadero, most of the trips are made between 12 pm – 4 pm. During this 5-hour window, demand accounts for 53% of trips from the whole day. Since these customers would likely be tourists, we can expect these stations to be fluctuating heavily with bike availability. The average lowest rental duration for this 5-hour time frame is 70 minutes, with the highest at 100 minutes. Looking at this information, it is clear that renters are taking it for more extended periods, especially since they are returning it to the same location. Potentially for 70 to 100 mins a day, there could be no bikes available, or renters are waiting until docking space opens up to park in the same station.
Now exploring the popular start stations, the highest traffic stations are around the Caltrain and Embarcadero. With the Caltrain stations, it would make sense for renters to exit a train and look to continue their journey to their office or home with a bike. The numbers of trips for Caltrain stations are significantly higher (56k – 73k), with the next being the Embarcadero (41k – 50k). Similarly, looking at the most popular return stations, we have the exact top 5 most popular locations with the start station results. However, the number of trips is almost double for the top return Caltrain station on Townsend at 4th at 92,000 trips. It is evident that customers are leaving work or home to take the Caltrain, and are coming from a variety of stations. Focusing on the most popular Caltrain start stations – Townsend at 4th and 330 Townsend, we see the highest demand is between 7 am – 9 am. Approximately ~ 54-55% of all trips in the entire day start between 7 am – 9 am. Shifting the focus to the most popular Caltrain return stations – Townsend at 4th and 330 Townsend, we see the highest demand is between 4 pm – 6 pm. With approximately ~ 50-54% of all trips in the entire day returned between 4 pm – 6 pm.
Recommendation
The best strategy for the company would be to repurpose the stations which have high bike availability and are not being used. 65 stations have 80% of their capacity in bikes available. These bikes can be moved into new additional stations focused around the Embarcadero area because the demand is higher for subscribers (68%) compared to customers (32%). Adding new stations in this area will accommodate subscribers who are more consistent with their rental usage. The average duration of all trips is 16 minutes, but the average trip duration for the top 10 most popular routes is 10 mins. The most popular locations have a shorter rental duration because the users are familiar with their start and return station. Providing more stations to support these groups would be most beneficial to the company and help solve the issue around bike shortages.
Dashboard Screenshots




data analysis
data visualization
data studio
dashboard
sql