We’re one week into our project with Benchmark, applying privacy-preserving techniques to city mobility data. These techniques allow you to see representation in a data set without compromising people’s privacy. This is especially needed in the mobility sector.
People rely more on cycling to travel post Covid
In the last few months people have been travelling around cities differently. Since Covid-19, more people are cycling and walking rather than using public transport. In response, city planners have rushed to accommodate for these changes. They have fast-tracked projects to widen pavements and improve cycling infrastructure that would normally have taken years.
There are inequalities in mobility and they’re not well measured
There are known inequalities in how Black communities access micro mobility. Yet there aren’t many metrics or datasets to help organisations and communities understand and measure these inequalities.
Access to micro mobility is unequal for a number of reasons. Benjamin Schneider thinks one factor preventing Black communities from accessing bike share schemes is the lack of information available. People aren’t aware of discount memberships and are concerned about liability fees if bikes are stolen. Bike share services like Uber’s Lime and Jump need to know how they underserve Black communities in their services to better meet their needs. Schneider also flags safety as a key concern, and not just from vehicle traffic.
Dr. Destiny Thomas speaks about how Black people are regularly policed, harassed, and killed in the built environment. This prevents Black communities from using more open and vulnerable transport options like cycling. Kevin Hylton describes mixed experiences as a cyclist of African-Carribean descent. When riding with a group of Black friends he suffers more micro-agressions and racist slurs than when riding with white cyclists.
City planners need to design spaces that prioritise protection from racism as well as vehicle traffic, to improve access to cycling. To address these inequalities and measure meaningful impact, we need ways to see representation in mobility data.
Mobility data is sensitive, and must be handled responsibly
Typically it has been hard to open data about mobility because it’s so sensitive. Last year Los Angeles Department of Transport demanded micro mobility services share anonymised data about customer’s journeys via the Mobility Data Specification (MDS) or lose their license to operate. Yet sharing this data impacts riders’ privacy and gives authorities the ability to track where citizens go. Even if you remove identifiers like names and addresses from the data you can still easily identify someone. In 2014 New York City Taxi and Limousine Commission released a dataset about taxi trips for an FOI request. Privacy researcher Andrew Tocker was able to re-identify the data and cross-reference it with paparazzi photographs to identify celebrity taxi trips.
Planners are making faster decisions about transport infrastructure in response to radical changes. It’s important for them to access data they need to support good decision making. Micro mobility providers have data that could be shared for public good. But this doesn’t have to be at the expense of people’s privacy.
Privacy-preserving techniques may give communities access to new datasets
Mobility data shouldn’t just be available to authorities. In the recent Reimagine our cities event, Javier Lopez from Red Hook Initiative spoke about how Black communities are constantly surveyed. But the data is extracted from communities and used exclusively by city planners. Lopez advocated for data to be shared with communities so they have agency and input on transport decisions and changes that impact them. More people could access this incredibly valuable data if organisations apply privacy-preserving techniques.
In the next blog-post, we’ll break down how to apply the randomised response technique to mobility data to see representation. As always, get in touch to share ideas or projects about equality in mobility data or privacy-preserving techniques.