How is B-Line Different Than Google?

How is B-Line Different Than Google?

One of the questions we get asked a lot when describing what B-Line does is: ‘How does B-Line’s location logging differ from what Google* can do?’

*the implication being that location history is what’s referred to

The simple answer is it’s a question of data granularity. Ok… but what does that mean?

When Google collects location information from you and me, it collects this information infrequently and with no concept of schedule or route. It just picks up points periodically when you launch an application and check for directions, or when you ask it to calculate a route for you.


As one can see from the map above, Google knows I went on a road trip in 2014 and visited a good part of the US. It knows what dates I was where and knows some information about travel within cities if I happened to use Google Maps while I had location services active on my phone, but it doesn’t know how I got from A to B, nor what my departure time was, nor where I stopped along the way for gas, what highway I took, etc. It knows some of the information, which is really valuable, but not the type of information necessary for infrastructure planning within cities. In essence, the data Google collects is patchy.

This patchy stuff Google has is really interesting information, in that it allows us to better understand patterns of inter-regional transportation (long distance trip making), something very poorly captured by tradition travel surveys. One could even make the point that this is terribly important when trying to understand sustainability of travel more broadly, as emissions related to inter-regional travel can account for 50% or more of all emissions related to transportation for urban dwellers who otherwise rely on transit and active modes (walking and cycling) for their day-to-day travel.

Unfortunately, this doesn’t help us understand travel within our cities for a variety of reasons. 

To begin with, Google knows where I live and work because I’ve told Google the address of these places. I volunteered this information to make my life easier and make queries faster. But despite having told Google (literally) hundreds of times that I want cycling directions, it often mistakes my cycling for driving. Or at other times doesn't know what to make of my travel:


This is not to say it doesn’t sometimes guess right, but even in those cases, when it gets the origin and destination right (especially if those are the home and work I provided and a familiar WiFi signature is present at both ends of that trip), it misses a whole lot of the additional detail along the way.

In the ‘history’ presented above, Google recognizes that I went from home to work, but it misses that I made a trip in between these two to buy bolts. And since I’m not a technically inclined person, this wasn’t a quick stop for bolts, it was 15 minutes of wandering around, not knowing what I needed.

One could say that it doesn’t matter that I made that additional trip, but that would be a fundamental misunderstanding of the motivation for collecting travel data within cities.

Taking a look at the total travel demand in Montreal for example, using data prepared by Zahabi et al. (2013), one finds that non-work and non-school purposes travel account for over 55% of all trips made, 36% of the total distance traveled and 43% of the total emissions generated. And this is a conservative estimate, as non-work, non-school trips are known to be under-reported in surveys such as Montreal’s self-reported household origin destination survey.

What this means is that it’s not enough to know where a person lives and works if one wants to plan transportation infrastructure for the region and work toward a less congested or more sustainable future. One also has to know how many other trips people make, where these trips are made and what modes were used to make them. If you’ve ever been stuck in traffic on a Saturday afternoon coming back from the mall, you know that traffic and vehicular movements aren’t just something that happens 7 - 9 AM and 4 – 7 PM Monday to Friday. That may be when network-wide conditions are worst of all, but lots happens outside that.

That being said, assuming one did not care about non-work, non-school trips and were to try and make the argument that peak travel is the only thing that matters (which completely disregards any question of transportation equity within cities, as well as assumes a travel time distribution that is inflexible and impossible to alter using policy instruments and road pricing), then there would still be the problem of not knowing the mode of transportation used by the majority of residents of a city using Google data, as well as not knowing what routes or specific streets are taken and by whom.

My travel last Tuesday according to Google (above). The first problem is that I did not walk and drive to York University for the 10th anniversary of the City Institute, I took the subway, followed by a bus. Google can’t know that without a higher frequency of location traces, so it just makes a guess. It also can’t know whether I took Bloor-Danforth across and up Spadina to grab a bus at Downsview or whether I took the Yonge line up to Finch and rode a bus west from there.

Answer: Downsview for the ride over, then, on the way back, I walked from York to the Wine Rack at Steeles and Dufferin (don’t judge me, I promised my significant other I would come back with a bottle of vino and that was the closest store that was open past 9PM on a Tuesday…).

What makes me all the more confused is that I took this rather long walk through the desolate space that is Steeles, periodically checking the GoogleMaps app for directions as I’m not familiar with the neighbourhood. Despite my having used a Google app multiple times to find the store, the traces don’t appear in my timeline and it instead looks like I spent 97 minutes getting home.


Is it important to capture everything? Well, that depends what you’re trying to better understand. To capture flows from region to region or get a rough idea of the connection between home and work locations, you don’t need to install B-Line. But to better understand how and when the citizens of your region are moving about, more detailed information is required. Capturing information on more trips, with higher spatial accuracy, and with mode of transportation and route to me seems like a better dataset to plan your infrastructure with.

Higher frequency and higher spatial accuracy location points that one can correlate with certain demographics (workers, students, men, women, youth, the elderly, etc.) means one can better understand the effect of varying levels of accessibility, traffic conditions and the built environment more generally on the choices individuals make. Without information on the paths people take and disaggregate records of trips made by different modes of transportation between different zones in a region, it is not possible to properly understand travel demand, and by association, plan for the future according to demographic and social trends.

So use the Google stuff to get a picture of high level regional flows, but leave the detailed data collection up to us!

Chris Harding, Chief Operating Officer, B-Line Inc.

Halifax Needs Smarter Transit Data

Halifax Needs Smarter Transit Data

The most recent release of Halifax Transit's Moving Forward Together Plan has generated an enormous amount of public discussion around the need for better public transit in the City of Halifax.  With declining annual ridership and local employers refusing to hire workers who rely solely on public transportation, the public is asking Halifax Transit to find new ways to design the city's transit system.  

Last week The Coast published an insightful article by It's More Than Buses outlining a number of challenges with the proposed route structure in the Moving Forward Together plan. In the article, Scott Edgar asks Halifax Transit to provide additional information regarding travel times and job accessibility on the proposed network before the plan is implemented. The article emphasizes the need for additional transportation data to ensure new investments in transit infrastructure will actually reduce commute times in the city. As Edgar explains,

"We know getting this data will be labour-intensive and will take a lot of time. But the stakes for the corridor routes are high. They will be the backbone of HRM’s transit system for decades to come, shaping the system’s growth for at least the next 30 years. Let’s take the time to get them right."

Understanding resident's travel behaviour is essential to designing efficient, reliable, and cost effective transportation systems. The Moving Forward Together plan, as it is currently presented, does not accurately reflect how Halifax residents actually move. The reason being is because Moving Forward Together was formulated using traditional methods of data collection such as: self-reporting surveys, transit pass sales, as well as manual vehicle and pedestrian counts. These methods of data collection are not ideal for transportation designers, as study participants are asked to estimate their daily commute patterns, while traffic counters are unaware of the origin, destination, or purpose of the commuters they are counting. As a result of these shortcomings, larger cities such as Singapore, San Francisco, Los Angeles, Montreal, and Toronto have begun to use mobile engagement strategies to better understand how residents actually move. 

Smartphone-based travel surveys are an easy and effective means of engaging residents in making their transit system more efficient.  Users simply download an app that passively collects their daily travel patterns, including their modes of transportation, eliminating the need for inaccurate and labour intensive self-reporting surveys.  In 2016, it is estimated that 68% of Canadians have access to smartphones, which is significant base of users to draw upon for high quality transportation data. Even though, Halifax has missed opportunities in the past to enhance its transportation system, residents are now requesting additional transit data to better understand where transit infrastructure is needed most.

The time has come to do away with the guesswork of traditional transit surveys by asking residents to share how they actually commute, rather than how they think they commute. Without modifications to the way transportation data is collected, unreliable service and declining transit ridership will continue for decades to come.