URBAN BIAS
I once listened to a podcast that outlined an example of natural bias in urban design. From snow plowing patterns in a small Swedish town to dummies for seat-belt testing, these were all designed around the lifestyle or physique of an average man. However with some data collection and analysis these gaps have been identified and systems have been changed to address these biases in our urban economy.
Given the common perception of there being a gender gap in the tech-world that dominates San Francisco, I took a look at it from the perspective of bike-sharing.
Collecting data from BayWheels for a 12-month period (November 2018 to October 2019), encompassing over 2 million rides, I first conducted a descriptive analysis to understand rider behavior and then a statistical analysis to dig into any potential gender gaps.
RIDER BEHAVIOR
ROUTES FOR TECHIES
First took a look at the nature of the trips - where are the users riding from and where to. Below is a graphical representation of the start and end points with the size and color representing daily trip volume.
South of Market (SOMA) and the Financial District (FiDi) are by far the top start and end points. This makes sense as this is where a lot of the tech office spaces are located. Mission is also a popular trip route as this is a nearby residential neighborhood.
For a better geographic representation of the above routes, the two business neighborhoods - SOMA and FiDi - are located next to each other and account for 61% of trips (yellow). When adding in the adjacent residential areas, this accounts for 86% of trips (green). This indicates how focused the user base is in this small portion of the city.
COMMUTER ORIENTED USAGE
Based on the insight that most trips are starting or ending in the business district, I next broke out hourly trip data as well as weekday vs. weekend behavior. It's clear that the majority of trips happen during commute hours, including both morning and evening rush hour. Weekend trips are overall minimal and more evenly distributed with a slow rise and decline, peaking mid afternoon.
Further breaking down the trip origins for weekday vs. weekend, the first graph below supports the idea that these trips are for commuting purposes towards business areas. However, looking at the second graph below, which represents weekend starting points, there are spikes in the Mission district as well as the Castro, which are more entertainment heavy neighborhoods with restaurants, bars and parks, aligning with expected weekend activities.
UNDERUTILIZED NEIGHBORHOODS
One final descriptive analysis I took a look at was the trip proportions in comparison to the proportion of docking stations in that neighborhood out of total stations. The next graph depicts how the two destinations - SoMa & Downtown are overutilized whereas Mission and Western Addition are underutilized in terms of ratio of trips indexing against ratio of stations.
The question of actual inventory spots per station is an additional factor here that was not considered, as well as distance between stations (stations per square mile) which may also influence the nature of a trip and thus total trip volume for that neighborhood.
OPPORTUNITIES
Based on the above descriptive analysis, initial opportunities identified include:
Capitalizing on weekend and non-commuter hours
Expanding beyond major business districts
The first point can be addressed through discounts or incentives sent to users through push notifications or marketing efforts towards non-users. The second point addresses a bit of a chicken and egg problem, but could be a way of tapping into the non-9-to-5 individuals who work in other neighborhoods or partner with the entertainment entities to offer promotions at the destination when using bike-share as the mode of transport.
INVENTORY GENDER GAP
GENDER SPLIT
Unlike popular belief, the population in San Francisco is split relatively evenly between men and women - 51% vs. 49% respectively. However when looking at bike-share usage it is highly skewed, with males accounting for 77% of the trips. Age distribution is similar between the two genders, so why the gap?
LOCATION DIFFERENCES
Where are women users riding from? To minimize the differences in docking stations per neighborhood as identified in the earlier analysis, trip per station is used as a benchmark against percentage of trips taken by women. Looking at the the top 10 high volume neighborhoods, there is a 70% negative correlation between trips per station and percent female trips, with the 3 neighborhoods ranking highest in female proportion coming in on the low end of trips per station. Though it's not the strongest correlation it could be an area worth exploring.
Neighborhoods with high proportion of female trips:
Bernal Heights
Noe Valley
Haight Ashbury
Marina
Western Addition
INVENTORY CHECK
Below is an inventory check by neighborhood, ranked by daily delta in bike stock, with the lowest inventory at the top. Comparing this against our above graph, the three neighborhoods identified above rank in the top 5 neighborhoods with lowest inventory. Women accounted for 28% of riders in these neighborhoods compared to 22% in the overall rider population, but they were among the least stocked. It is difficult to identify whether there may be other factors at play here, but it could be a good starting point to increase female ridership.
RECOMMENDATIONS
After digging into these inventory gaps, some additional recommendations include:
Implementing marketing strategy oriented towards women
Increasing bike inventory in Western Addition to start
Address the first point by surveying women to better understand their perspective or hesitancy towards using bike-share. Results of this should influence the ultimate marketing messaging. For inventory management, Western Addition is currently where the largest gap lies in terms of inventory, but there may be other areas where the current system is missing a key user base, so additional analysis is recommended.
FUTURE OF MICROMOBILITY
FURTHER ANALYSIS
With a focus on San Francisco, some interesting next steps include:
Researching impact of topography, particularly hills
Analyzing the relationship with increase in bike lanes
Discovering the impact of car-free Market street
San Francisco is a very unique urban bubble. However many of these current approaches can be used to initialize optimal plans in other cities. Although bike-sharing just touches the surface in understanding the larger micromobility landscape, there are learnings to be applied in the greater scope of urban transportation. As data is continuously gathered to improve and spread low emission transport usage within cities, it will become increasingly important to build systems accessible to all residents, not just the most likely user.
Comments