Edit 2014-05-31: As this post appears to be getting a lot of search engine traffic, I'd like to point out that the functionality I'm describing here is now available much more comfortably as part of Strava itself. I guess I was not to only person intrigued by the concept.
Most people would not think of Strava as a planning tool. Sure, it's a great performance and training tracker and provides lots of opportunities to compete with fellow cyclists and yourself, but its use for route planning is less obvious.
The basic problem is this: You want to plan a ride in area that you don't know well. There are many roads, but how can you tell apart the heavily trafficked, narrow highway from the lovely scenic back road? There already are a number of tool that can help, but they all have some shortcomings. For the amount of traffic on a given road, one can look up traffic count data from state or local departments of transportation (e.g. New York State). Depending on your location, this works reasonably well for major roads, but the small back roads that are often the nicest to ride on often lack count data. Google Street View can give you a good general impression of a road -- how wide is it, does it have a shoulder, are there big potholes? -- but it can be tedious to check longer stretches of a road. OpenStreetMap allows everyone to add useful information to their geographic database, like the existence of shoulders, the smoothness of the road, or its speed limit; but so far a lot of those features are not displayed on the map and the data especially in North America is very incomplete.
 |
| Probably a nice road for cycling -- unless the road is only this empty on early Sunday mornings ... (Screenshot from Google Street View) |
So wouldn't it be great if you could just ask a fellow cyclist what she thinks of the road in question? Or, as different cyclists have different preferences, if you could ask a hundred cyclists? Or a thousand? Well, why not just do that: Hundreds of thousands of athletes log their bike rides on the Strava website, and fortunately Strava provides an API that allows other applications to access that data. The folks at raceshape.com offer a number of different analysis tools, but for our purposes the "global heatmap" is the one that's most useful.
What the heatmap does is to basically accumulate all the GPS data from Strava and overlay it on a map, with the color of the tracks representing the number of times a road has been ridden on. If you look at Montreal, for example, you can very clearly see some of the most popular cycling routes: the Estacade across the St. Lawrence, the Lachine Canal bike path, or the various ways to climb and descend Mont Royal.
 |
| Screenshot from raceshape.com |
Just like the approaches described above, there are limitations to the global heatmap for ride planning as well: Less populated areas often don't have sufficient data to be displayed on the map. In addition, it's important to keep in mind that Strava is mostly used by road cyclists, evidenced by the fact that for example the very popular bike path along the Chambly Canal doesn't even appear on the map -- presumably because it has a gravel surface.
 |
| Some traces of activity on the two highways along the Richelieu River, but nothing on the bike path (Screenshot from raceshape.com) |
Update: As Andy and Brian have pointed out in the comments, the data used for the global heatmap is more limited than I thought: Instead of representing all or even a large portion of Strava ride data, it is only a small subset of those Strava users who explicitly have allowed the raceshape website to access their data. And at the moment there isn't even a link on the raceshape homepage to do that. If you would like to have your data included and get your own personal heatmap, use this link.
