Strava moves into 'big data' - London & Glasgow already signed up to find out where cyclists ride

Company behind apps that track running & riding targets local authorities & advocacy groups with Strava Metro

by Simon_MacMichael   May 7, 2014  

Strava Metro London UK

Strava has moved into the 'big data' game with the launch of Strava Metro, which it says gives data providing “ground truth” on where people ride bikes or go running – and it is licensing the results to city authorities and advocacy groups, including in London and Glasgow.

The San Francisco, California-based company, developer of the smartphones apps and website that allow users worldwide to track their rides and runs, says that “millions of GPS-tracked activities are uploaded to Strava every week from around the globe.

“In denser metro areas, nearly one-half of these are commutes. These activities create billions of data points that, when aggregated, enable deep analysis and understanding of real-world cycling and pedestrian route preferences.”

The popularity of using Strava on main commuter routes can clearly be seen on the map of London accompanying this article – you can find a bigger version here on the Bicycleretailer.com website – with strong levels of usage on roads such as the Embankment.

Making the data available to local transportation authorities or advocacy groups can help identify where demand for cycling, for example – and thereby the need for safe infrastructure – is strongest.

According to the company, “Strava Metro’s mission is to produce state-of-the-art spatial data products and services to make cycling, running and walking in cities better.

“Using Strava Metro, departments of transportation and city planners, as well as advocacy groups and corporations, can make informed and effective decisions when planning, maintaining, and upgrading cycling and pedestrian corridors.”

Clearly, there are bound to be privacy concerns with such a service – we’ve reported in the past, for example, concerns that thieves use rides uploaded to Strava to target where people who own high-end bikes live.

Strava has sought to allay such worries by emphasising that Strava Metro “processes the data to remove all personal information linked to the user and structures it for compatibility with classic geographical information systems (GIS) environments.”

It adds: “Strava Metro tools enable DOTs and advocacy groups to do detailed analyses and glean insights into cycling and running patterns dissected by time of day, day of week, season and local geography.

“Advocacy organisations and the general public can now access high-resolution heatmap visualisations of the data free of charge at Strava Labs.”

You can find those heatmaps here.

“Organisations seeking deeper insight and analysis can license Strava Metro data and tools for use with geographic information systems (GIS) mapping software. Licensing costs are based on the number of Strava members in the requested geographic area.”

There is an inquiry form for anyone wanting to find out more information.

The company’s co-founder and president Michael Horvath said: "Bicycling safety is a top concern to our members worldwide, especially when they're riding through metropolitan areas with a high concentration of motor vehicle traffic.

"Strava Metro delivers an innovative way for us to serve Strava members and non-members alike by helping to make their daily commutes and weekend rides smoother and safer," he added.

Pricing of the licensing of the data will depend on the number of Strava users in the area concerned.

According to a blog post by Reed Albergotti in The Wall Street Journal, the first local authority to sign up is the department of transportation for Oregon, which will pay $20,000 to license it for 12 months to analyse usage in Portland.

A policy analyst working for that body, Margi Bradway, said: “We’re dipping our toe into the idea of big data with this project.”

Jennifer Dill, who is a professor at Portland State University’s Nohad A. Toulan School of Urban Studies and Planning, said: “Right now, there’s no data. We don’t know where people ride bikes. Just knowing where the cyclists are is a start.”

Other customers are located in London, Glasgow, Orlando in Florida, and Alpine Shire in Victoria, Australia, although the actual bodies to have licensed the data have not been reported.

46 user comments

Latest 30 commentsNewest firstBest ratedAll

Strava is the devils own... Turning the wonderful act of riding alone or with mates into mindless willy waving.

If any policy or data is used from that source, then it's study and use is to highlight the preferences and activities of mindless willy wavers in society.

(God I'm getting old and grumpy)

posted by MisterT [6 posts]
8th May 2014 - 7:59

17 Likes

First let me say I log everything on Strava which I think is a bit sad, hay ho so is wearing tights in you mid forties Worried ... Joking aside, sharing of people's personal data with corporate anyone, always makes me worry. Put in the hands of say the Daily Nail ha ha and I'm sure the information could be used against riders, also will it mean the less used routes (by strava riders) will be neglected. Perhaps it would be better to make non car infrastructure a priority. I know I'm dreaming about that. I honestly don't see most authorities using the data for good.... I only seem to hear how evil cyclists are these days. It's not convenient to admit we take cars off the road, safety for pedestrians and the environment. This will be ignored to point out riders going on pavement or the wrong way on one way ect.

Bloody cyclists!!! Devil

HMCC

Beefy's picture

posted by Beefy [174 posts]
8th May 2014 - 8:09

34 Likes

Local authorities should be providing cycling facilities useable by those covered by the snappy phrase "from Eight to Eighty".

Do eight year old kids use Strava en masse?

Do eighty year olds use Strava en masse?

Can we trust local authorities to use this information from Strava to cater for those groups?

Local authorities will use this information to make lazy decisions and back them up when they install piss poor facilities.

posted by farrell [1580 posts]
8th May 2014 - 8:10

24 Likes

Fascinating stuff. Given that the data will include time as well as location, I don't think it matters that the majority of the data might be distorted in favour of segment-chasers - by extracting the information of the slowest 10% at prime commuting times, I'd imagine authorities will achieve a fairly accurate snapshot.

dafyddp's picture

posted by dafyddp [181 posts]
8th May 2014 - 8:29

18 Likes

I live 20 mile outside of Glasgow although there are many roads that are favoured by cyclists, some of the best roads are less popular.

I'm also not fond of The National Cycle Path with it's tree roots poking through the concrete, excess gravel, the sewage plant and the remnants of the Buckfast fuelled nights left nearby.

stenmeister's picture

posted by stenmeister [76 posts]
8th May 2014 - 8:34

19 Likes

I've just wasted five minutes looking at that heat map again. Interesting to see where people are going and what sort of level of cycling warrants what depth of heat... looking at my folks house, I know I will be the only person using their driveway, and despite two years of strava fun, it barely resonates on the map.

Quite fun to look at though.

posted by Jimmy Ray Will [354 posts]
8th May 2014 - 9:07

20 Likes

My various commuting routes all look popular on the map which seems about right to me.

Interesting that Herne hill velodrome and the crystal palace crits are shown, I guess these activities occur around commuting time.

posted by Scoob_84 [233 posts]
8th May 2014 - 10:09

17 Likes

Excellent another 'big data' story. I look forward to someone trying to sell me totally irrelevant products whilst I am cycling along trying to concentrate on the roads. Maybe the ambulance chasers will be the first ones to take up the new advertising opportunity...

posted by rdanderson [4 posts]
8th May 2014 - 10:35

20 Likes

nowasps wrote:
I only record my longer "lycra" rides on Strava. All my commuting, shopping and city riding is done without.

Exactly. Counting up where the vast majority of people using bicycles - and they do NOT have Strava - means spending the money on doing proper cycle traffic counts.

posted by ChairRDRF [169 posts]
8th May 2014 - 10:40

17 Likes

joemmo wrote:
mrmo wrote:
I think it is a no brainer why this information could be useful, my only caveat, how representative of ALL cyclists is Strava?

Probably better than nothing but unlikely to capture the non competitive and utility riders who will benefit most from better infrastructure.

Actually, quite a lot worse than nothing - it would give an entirely false and misleading impression of cycling as a whole to use data from a small coterie of speed demon vehicular cyclists, and decisions made on such data would almost certainly be bad decisions. I do hope it is not true that any London local authority has bought this crap.

posted by Paul M [325 posts]
8th May 2014 - 10:54

20 Likes

Garbage in, garbage out. You only have to look at Strava's "heat map" across northern Europe to see what garbage it produces. It suggests that the UK is a more active cycling nation than the Netherlands which is palpable nonsense. I even suggests that there is more cycling in the UK - by a wide margin - than in France, and even at the level of weekend sports riding that is a nonsense as anyone who has driven a French road can attest.

Lord help us if this rubbish gets used to inform local authority investment decisions.

posted by Paul M [325 posts]
8th May 2014 - 10:58

16 Likes

So are people saying that commuters who strava to work pick different routes through London than commuters who dont?

posted by Scoob_84 [233 posts]
8th May 2014 - 11:06

22 Likes

Well, this is just fantastic for quite a few reasons:
1) I commute a lot from London City to Essex. This was a great way to find out if there are other "secret" routes I have not considered.
2) It quickly points out that the main roads in outer London are heavily used by bikes and the back roads are not. I.e. where to focus the investment. I'm talking about examples like the Bow interchange/flyover and other routes near Canary Wharf.

This is the case that a picture is better than a thousand words.

Now lets use this to get the cycling routes sorted out/maintained.

posted by schocca [4 posts]
8th May 2014 - 11:36

25 Likes

Great that Strava are pushing these towards councils - HIGHLY COMMENDED !!!

posted by yenrod [104 posts]
8th May 2014 - 13:43

20 Likes

Paul M wrote:
Garbage in, garbage out. You only have to look at Strava's "heat map" across northern Europe to see what garbage it produces. It suggests that the UK is a more active cycling nation than the Netherlands which is palpable nonsense. I even suggests that there is more cycling in the UK - by a wide margin - than in France, and even at the level of weekend sports riding that is a nonsense as anyone who has driven a French road can attest.

Lord help us if this rubbish gets used to inform local authority investment decisions.

Amen to that.

It's a heat map of MAMILs not cyclists. End of.

Really, though?

posted by workhard [389 posts]
8th May 2014 - 14:02

15 Likes

I disagree

posted by Scoob_84 [233 posts]
8th May 2014 - 14:19

13 Likes

schocca wrote:

2) It quickly points out that the main roads in outer London are heavily used by bikes and the back roads are not. I.e. where to focus the investment. I'm talking about examples like the Bow interchange/flyover and other routes near Canary Wharf.

Sorry, but it won't show that at all.
You only have to think about the percentage of people you see on your commute who you think might be using Strava to see the problem with this approach.

posted by Chuck [435 posts]
8th May 2014 - 14:25

18 Likes

In Cambridge, some routes that are known for commuter cycling congestion are very, very faint. Whilst routes that are known to have almost no commuting are very, very bright.

posted by Al__S [646 posts]
8th May 2014 - 15:06

11 Likes

Chuck wrote:
schocca wrote:

2) It quickly points out that the main roads in outer London are heavily used by bikes and the back roads are not. I.e. where to focus the investment. I'm talking about examples like the Bow interchange/flyover and other routes near Canary Wharf.

Sorry, but it won't show that at all.
You only have to think about the percentage of people you see on your commute who you think might be using Strava to see the problem with this approach.

Let's be clear here - if you want to traverse a decent distance in London, you are not taking a back route, it's just too convoluted. So I'm happy to take the Strava data as a proxy on what is going on. Is there any other stats that are available? If so, happy to get a pointer.

For East London - the A12, A406 and the M25 are effectively massive firewalls for travelling East/West. You only have a few routes to get across these without becoming a Jam tart. So the Strava data is pretty relevant.
- Example - Commercial Road (City to A13). This is not a signed bike route, but you can see a medium level of bike traffic using it via the Strava data. I use it as well to access the A13 CS3 beyond Canary Wharf. It's nasty and intimidating, but significantly faster than using the "pedestrian friendly" CS3 between City and Canary Wharf.

posted by schocca [4 posts]
8th May 2014 - 15:22

20 Likes

Quince wrote:
I can't imagine what big corporations could really DO with lots of cycle data. But mixing big corporations with big data has a rather ominous ring to it.

So you don't want new bike shops to pop up on heavily used routes?

posted by Initialised [197 posts]
8th May 2014 - 16:42

18 Likes

Yes that's true but it's not "End of".

The fact that it is not representative of the entire cycling population is not a problem as long as you know what segment the sample represents.

As I said above, if used in conjunction with 'small data' then this sort of thing can be useful.

By using more detailed and representative information such as traffic surveys you can effectively calibrate the data and make adjustments for such things as representativeness.

It's done all the time. The only concern here is that the people using this data know that they need to do it.

I know what I'm talking about on this, as I run a research agency.

abudhabiChris's picture

posted by abudhabiChris [569 posts]
8th May 2014 - 16:52

29 Likes

mesagsx wrote:
Before everyone gets too animated about whether a Strava user is a "real" cyclist ... It's pretty straight-forward for Strava to analyze the data and separate the commuters out from the roadies & others, based on their speed, pattern of acceleration and time of day.

What you're seeing here is the untreated version of that data - if you want the breakdown and more detailed analysis, I'm sure it's available from Strava for a price Smile

Analyzing vast swathes of data is never perfect, but this is a pretty snazzy source of data IMHO, and a really important step towards understanding where we all ride and what might make it easier.

Are you sure about this though?

Perhaps they can filter the data they have, but they can't put back in data that was never there in the first place. Its always going to be a partial and biased sample, as large numbers of cyclists will not be using Strava to begin with (I don't - no disrespect to those who find it useful or just amusing, but I have no interest in it).

Unless you can be sure its entirely random whether someone registers with Strava or not then how can they know what data they are missing?

posted by FluffyKittenofT... [753 posts]
8th May 2014 - 22:46

13 Likes

abudhabiChris wrote:
Yes that's true but it's not "End of".

The fact that it is not representative of the entire cycling population is not a problem as long as you know what segment the sample represents.

As I said above, if used in conjunction with 'small data' then this sort of thing can be useful.

By using more detailed and representative information such as traffic surveys you can effectively calibrate the data and make adjustments for such things as representativeness.

It's done all the time. The only concern here is that the people using this data know that they need to do it.

I know what I'm talking about on this, as I run a research agency.

I guess this is partly an answer to what I just asked!
I remain unconvinced that the necessary information to properly 'calibrate' this data is going to be available though.

The danger is the 'unknown unknowns' to invoke Donald Rumsfeld!

I mean, what if you do your traffic survey and then discover that what it shows is that there is no consistent correlation at all between Strava traffic and non-Strava cyclists?

posted by FluffyKittenofT... [753 posts]
8th May 2014 - 22:52

12 Likes

I love that we are all speculating on what the data can or cannot show, without access to that data, or any demographics of Strava users what so ever!

What is likely, is that this probably represents some of the largest and most detailed datasets on riding ever collected, and there will undoubtedly be value that can be gleaned from this. I would be shocked if any other studies will have been able to GPS track a group of cyclists on anything like the scale Strava has.

posted by giobox [342 posts]
9th May 2014 - 1:11

10 Likes

giobox wrote:
I love that we are all speculating on what the data can or cannot show, without access to that data, or any demographics of Strava users what so ever!

I'm not speculating. I know for a fact that on the publcily available activity heat map- the data for councils is just this, with some filters- some of the busiest cycling routes locally (ones that get actual traffic jams of cyclists) show almost no activity, whilst roads with known small numbers (from council counting) are very bright.

posted by Al__S [646 posts]
9th May 2014 - 6:09

11 Likes

Al__S wrote:
giobox wrote:
I love that we are all speculating on what the data can or cannot show, without access to that data, or any demographics of Strava users what so ever!

I'm not speculating. I know for a fact that on the publcily available activity heat map- the data for councils is just this, with some filters- some of the busiest cycling routes locally (ones that get actual traffic jams of cyclists) show almost no activity, whilst roads with known small numbers (from council counting) are very bright.

You talking about London here or somewhere else? Because the commuting heat map shown above seems pretty accurate to me.

College road in Dulwich is used by all types of riders on the morning commute and is represented by a big thick orange line. It's also not somewhere that Lycra louts generally go on their weekend jollys.

posted by Scoob_84 [233 posts]
9th May 2014 - 8:35

8 Likes

I think the data will be useful (certainly in London), the ability to select times will show some interesting patterns.
For example it may show that a lot of cyclists are not using a crap cycling facility and staying with the road instead, showing that the facility needs improvement.

Most of the utility cyclist I know in London also use Strava, not just the people going on training rides.
Strava users will tend to be younger (but that is always the case with new technology), but younger riders do make up the majority on London's roads.

posted by thereverent [322 posts]
9th May 2014 - 9:13

8 Likes

If it shows where cyclists actually ride - direct main road routes - rather than where TfL hopes they ride - hopelessly indirect back street 'quietways' - then it will have achieved something, encouraging investment where cycling is already happening.

Space Syntax has also done independent research showing where cyclists actually ride:

http://www.spacesyntax.com/project/encouraging-cycling-in-central-london/

And surprise surprise, it's NOT on quiet back streets. Space for cycling needs to be provided on the main roads where most cyclists ride. If that includes the main road I ride along on Sunday morning, so much the better.

drmatthewhardy's picture

posted by drmatthewhardy [417 posts]
9th May 2014 - 18:57

8 Likes

Quote:
College road in Dulwich is used by all types of riders on the morning commute and is represented by a big thick orange line. It's also not somewhere that Lycra louts generally go on their weekend jollys.

Actually it is - a huge number of S. London based cyclists ride up that to meet at Crystal Palace on a weekend and also ride along that road to access Herne Hill Velodrome.

Speculating on what the data does or doesn't show isn't particularly helpful. I'd guess that this is the first step in Strava making their data available and this is just a massive data dump to attract attention (and therefore ££) from potential buyers. In it's raw form it's probably not *that* useful - after all I'd be willing to bet that very few casual commuters use Strava - but I reckon if you combine Strava with the heatmaps that have already been done on the Barclays Cycle Hire scheme, then work that in with traffic cameras, known accident blackspots, mainline stations, the congestion charge zone etc, it could actually be very useful.

posted by crazy-legs [568 posts]
9th May 2014 - 19:19

2 Likes

It looks Strava released a new FAQ about the data: http://metro.strava.com/faq/ . I think before I make any large leaps about what this data represents i might wait to actually see what the data is. To me it looks like an amazing opportunity to get ALL cyclists represented. I mean you can say Strava is only racers or you join and start to make your voice heard.

Ride on.

posted by Commutes_FTW [1 posts]
13th May 2014 - 16:43

2 Likes