Data Science with Madeleine Bonsma-Fisher

One critical component of road safety advocacy is the collection and interpretation of data such as bike counts, collisions, and economic activity. It’s also important to compare advocacy efforts in other cities. To dive deeper on both fronts, I spoke with Madeleine Bonsma-Fisher on October 11, 2023 about her passion for data science, advocacy with Bike Ottawa, and post-doctoral research on cycling stress levels in Toronto.

RZ: Walk me through how your passion for data and bikes came about.
MBF: When I was studying physics, I had a few research projects where I had to do some coding and got hooked. After those early experiences programming, I knew that's what I wanted to do along with working with data, pulling something interesting from it, visualizing it, and looking at it from different angles. I decided to apply that to biology, so I did a Biophysics Masters and PhD. I still love biology. It's very interesting and I have no regrets.

The whole time this is happening, I started undergrad, brought a bike along, and biked to and from my jobs or classes. What happens to most people who bike is you get more frustrated over time with how you’re not meant to be there, or your safety is second to traffic flow. The real turning point came when I started to learn about the advocacy community and things people were thinking would help. I learned about Vision Zero and thought it would be great if we could do this. From then on, I was always looking around and thinking how come I didn't think of that way before. We assumed the way we built our streets and cities is just the way it is, but there are all these people talking about the potential that we could do things differently. Bike Ottawa was the start, but there was also the City of Toronto’s Vision Zero data challenge (in 2018). Do you remember this?

Taking photos with large trucks are an effective way to show how dangerous they are on our streets

RZ: Vaguely. What more can you tell me about the Vision Zero Data Challenge?
MBF: At the time, I was involved with U of T Coders which was mostly grad students teaching each other programming things related to their research and doing data projects on the side. Around that time, someone introduced me to CivicTechTO and I went to a few meetings. The City put out this call for the Vision Zero challenge; looking for data analysis or visualization projects relating to Vision Zero goals. With a couple of friends, I started looking at the city's collision data which was the first time I did a data related thing that had to do with biking.

I remember Jennifer Keesmaat tweeted something about road safety which got all kinds of replies on Twitter. People were saying “it's always those bikes” running the stop light and doing all this unsafe stuff. I replied, this is not true. I analyzed collision data and found in bike-car collisions, the vehicle is at fault over half the time, so this perception that bikes are breaking the rules is not true based on police data which is not the most unbiased data set or even complete.

The next time I started bringing the two together was with Bike Ottawa. They have a Data Working Group which is volunteers with some interest in data or some skill related to data science. Whatever we wanted to do with data and biking, we could just do some little projects.

RZ: What kind of data projects did you work on during your time with Bike Ottawa?
MBF: What I was expecting to work with was stuff like the city's automatic bike count data – which I did a little bit with – but then I learned about this Strava data set Bike Ottawa had access to. It started as a fitness app but is now about general tracking of your outdoor activities. Strava was selling this data which the City of Ottawa paid for two years’ worth in 2016, but then they started making it free. You had to apply if you were an advocacy organization, research group, municipality, or government. When I learned about this, I got so excited and was immediately interested in learning how biking changed during the pandemic with Strava data.

The data shows hourly bike counts on every road and path in the entire city which is a very detailed data set. The first thing I did was look at the change between 2020 and 2019 (or before and after the pandemic). What we found interesting was you could clearly see on the map people were cycling less in downtown and more on residential roads and outer-lying areas relative to 2019, but all the City's automatic counters showed either the rates hadn’t changed or gone down. When we looked at where the counters were, we realized they captured popular commuting routes which is the opposite of this bike boom. The story we were hearing about during the pandemic looked like people were biking less, but that's because they were only looking at a subset of routes and trip types with automatic counters.

RZ: I heard there was a study on cycling stress levels in Ottawa. Were you involved with that?
MBF: No. That happened quite a few years ago (in 2018). Bike Ottawa got some funding to develop this ecosystem of maps which is very impressive. One of them is a Level of Traffic Stress map which uses OpenStreetMap and tries to either use the features annotated in the map (e.g. number of lanes, speed limit) or guess what they're going to be based on defaults, and then assign a cycling stress level to every road. I did end up working with Level of Traffic Stress in my current research.

RZ: Having biked in both Ottawa and Toronto, how you would compare the two cities?
MBF: I think it's hard to answer this because things have changed over time in both places and my own life situation has changed when we moved to Ottawa. I had a newborn and my risk calculation was a bit different than when I was biking in Toronto. For the places I tended to go, Ottawa has well-connected trails that are better connected than Toronto's trails, and generally worse on-street infrastructure than Toronto.

I was joking recently that Ottawa has both the best and worst bike infrastructure I've ever seen. One example that’s amazing is you’re coming west along Somerset towards the Rideau Canal on the east side. The road you’re on is higher than the canal. In Toronto, what would happen is the road would end, there would be a staircase, and nobody would bother connecting the path along the canal with the road. But in this case, there was a beautiful cloverleaf ramp that went down to the path so you could bike straight down it. But the path itself is narrow, covered in potholes, and has too many pedestrians and cyclists which is not great. My sense is there are many committed people who bike in Ottawa, but there are more general casual cyclists in Toronto. If you're on a downtown street in Toronto, there's just a lot of people out biking, so there is a little more of this herd feeling which always made me feel you belong and there’s safety in numbers.

RZ: How was your experience riding on Queen Elizabeth Driveway which Mayor Sutcliffe opposed?
MBF: There’s a path beside the road which – on a nice day – is full of people walking, pushing strollers, and biking; so there really is a space issue. The path is not wide enough to accommodate people who want to use it and none of the paths are wide enough to ride side by side even under less crowded conditions. A few times, I would go to the farmers market with a friend, and we could ride on QED side by side in our cargo bikes and just chat. Our kids could wave to each other and having more space was really nice.

There was a similar vibe with ActiveTO. Maybe not exactly since Lake Shore is an important road for driving – I don’t know that it has to be – but Queen Elizabeth Driveway was built as a recreational or scenic route and not meant to be an arterial road.

RZ: Which components are you planning to incorporate into your post-doctoral research on cycling stress levels in Toronto?
MBF: The basic idea is you use the Level of Traffic Stress rating framework developed about ten years ago in the United States. It's based on surveys which asked people where they would be willing to bike, which types of infrastructure, and where they bike. What they found is almost nobody who was willing to bike is willing to bike in the same lane as a car or even next to fast-moving cars on a painted bike lane. The levels go from one to four where one is fully separated from cars – either a protected cycle track, an off-road path, or a calm local street with low speeds and traffic volumes – which would be considered appropriate for anyone, even children. Level two is similar but could have a slightly higher speed limit or a buffered bike lane. Levels three and four have multiple traffic lanes, fast-moving traffic, and either a painted bike lane or nothing.

What the group did was they created this map for Toronto and rated every road and path. They looked at how many destinations you can access from various starting points by using only low stress roads. Adults from surveys don't want to bike on high stress roads, so let’s assume nobody will bike on them. What they found was the number of jobs you could access within a 30-minute bike ride on a low stress network was very strongly correlated with the likelihood of cycle commuting, so this is evidence that this rating system is reflecting people's behaviour. That’s the framework for everything we’re doing.

Here's a video of Madeleine explaining her cycling research in greater detail at the U of T Data Sciences Institute Research Day on September 27, 2023.

RZ: When do you plan to have your findings presented to the public?
MBF: Good question. Those papers that I just talked about came out in 2019 and 2021 before I joined this research group. Since then, a PhD student in our group made an optimization model to use the number of jobs you can access on the low stress network to try to maximize that with new infrastructure projects that are making high stress routes low stress. The pre-print of that paper is online.

The layer I've added on to it is thinking about the different choices in the optimization process, what you're prioritizing, and thinking about spatial and distributional equity in terms of who needs infrastructure and who gets infrastructure. We’re writing that up right now, so it will hopefully come out sometime in the next few months or half a year.

RZ: You and your family did a recent bike-packing trip around the Lake Erie area. How did you prepare for this trek with a kid in tow?
MBF: The big secret for us is an electric cargo bike which takes off the edge in terms of needing to pack super light. Last year, we did a similar Waterfront trip from Burlington to Brockville, and we rented two bikes from Happy Fiets. Since we already have an electric cargo bike this time, we rented the second e-bike from Robin for my husband to ride. My parents – who were also on the trip with us – did a lot of the planning in terms of the route and where we were going to stay, so I don't think we would have had the bravery to do this without them (and vice-versa). Most of the planning was done very far in advance because you must book camp sites in provincial parks five months ahead of time. I think the electric cargo bike takes this trip from a difficult character-building adventure to something pleasant and enjoyable.

We rented a second battery for one of the bikes which has 2 battery slots. We don't own a second battery, but with two batteries, the range is about 250 kilometers. Even then, we camped at a site every night which had electricity, so we could just plug it in on the campsite.

RZ: Tell me about your World Park(ing) Day installation from September 15.
MBF: About a year ago, I posted a tweet asking what you would put in a parking spot if you could put in anything you wanted. A lot of people replied and then somebody sent me the link to World Park(ing) Day. It’s not organized specifically – being just a website and a manual – but the third Friday in September is Park(ing) Day. The idea is you take a metered parking spot and you put something else there.

I think the original idea was like a tiny park in which you replace a parking spot with a bench, a tree, and some astroturf and just think about the space we give up for parking in our cities. Since then, people have gone more artsy. They've done little art installations and information booths on urbanism. I wanted this to be low key and just a little thing we were going to do with some neighbours. People brought some potted plants, lawn chairs, and a folding table. I brought a little plastic slide and some kids toys so it resembled a mini backyard patio. We had a couple of people from the neighborhood come and hang out for a couple of hours, and it was a good time.

RZ: What advice or resources would you share with road safety advocates looking to dive deeper into data analysis or data science?
MBF: I feel there's so much yet to be done in this space with data and quantitative stuff, while I'm always learning about cool things that people are doing related to this. I have a Twitter thread pinned that shows cool maps I've come across that display data about the function of an area including how many destinations you can access within a 15-minute walk and how many residents in this city live within a 15-minute walk of everything they need. As for getting involved in doing some of this stuff, advocacy organizations such as Bike Ottawa and Cycle Toronto are a great place to start. They all have these little projects going on and anybody can come in. You can pitch your ideas to others and do something if you have an idea.

From poking around the active transportation research world in the last few weeks I've learned about some very neat resources that I want to share with you.

✨Cool websites, maps, and crowdsourced data platforms for advocates, researchers, and practitioners!✨

A 🧵

— Madeleine Bonsma-Fisher (@mbonsma) November 6, 2022

UPDATE 2024/01/10: The research paper Madeleine Bonsma-Fisher was working on can now be accessed at this link & this interview ended up being the most read post on Two Wheeled Politics in 2023.