Strava heat map oddities and insights

[lordofthemark]

You can compare 2014 to 2015.

http://labs.strava.com/heatmap/2014-2015.html#13/-76.99914/38.90312/gray/bike

1. There is a big increase in cycling on North Hampton near Park Center. It is not just me, because Park Center Drive itself does not even show up as being ridden. I cannot see any particular closures on other routes that would cause it. The new bike lanes on a road that was already relatively bike friendly, seem to have made a big difference. Similarly either the Wheeler lanes had an impact or Port City Brewing has gotten more popular. Potomac Avenue has gotten busier (new development) but the advisory lanes in Potomac Greens seem to have made no difference (at least to Strava users)

2. You can really see the impact of the road closures at NVCC Alexandria, and where people have detoured to.

3. Some folks are already riding through Army Navy Golf Club. About as many as going through Arlington National Cemetary. How is this possible? I mean I would expect people riding in semilegal places, like the riverwalk in the Navy Yard, or Washington Harbour, but I thought the Country Club was seriously off limits?

4. A lot more people took the airport detour than the Eads Street detour to avoid the mulch.

5. Not much if any increase in riding within Tysons, but a big boost in riding from McLean to the McLean metro station. Take note, FABB.

6. Big increases on the streets of DC, in the places you would expect. Some of course might be increased Strava market penetration.

Topic

[runbike]

You can see a definite deepening of color along Wilson Blvd. (west of Ballston) where the new bike lanes were put in mid-2015.

[scoot]

Without knowing the color scale, it’s tough to draw many conclusions from this visualization.

Some of the increases are certainly expected due to new or improved infrastructure:
e.g. Pershing Drive heading toward Route 50 trail

Others seem to have occurred without significant infrastructure changes:
e.g. Beauregard between Duke and Armistead

Other spots appear to have lost a lot of riders:
S Van Dorn between Duke and Pickett (I’m surprised to see so many riders here…)
S Pickett St

Also, who’s been riding up and down the southern half of 395? 😮

[jabberwocky]

Hah. My office moved at the beginning of 2015, and my commute is so lightly used by bikes that I can see a noticeable difference between the two solely because of it. My 2014 route dropped in use significantly and my 2015 picked up a lot.

[dasgeh]

It would be useful to see which routes increased more than the average.

[consularrider]

Are you able to filter for just bikes? Otherwise I think the Army-Navy users may be runners.

[AlexandriaBiker]

@consularrider 131498 wrote:

Are you able to filter for just bikes? Otherwise I think the Army-Navy users may be runners.

The down arrow next to 2014 vs 2015 gives you options to show bikes, runners or both. As presented only bike data is shown.

[Tim Kelley]

@lordofthemark 131461 wrote:

3. Some folks are already riding through Army Navy Golf Club. About as many as going through Arlington National Cemetary. How is this possible? I mean I would expect people riding in semilegal places, like the riverwalk in the Navy Yard, or Washington Harbour, but I thought the Country Club was seriously off limits?

The gates/roads are open–there is no one to stop you from riding through.

[Raymo853]

I have also noticed some very odd mistakes in the heatmap. The heatmap is generated in a series of tiles then stitched together to produce a continuous image. Sometimes something odd happens with a single tile, and the there are very distinct changes along the edge of the tile. A close by example is Cedarville State forest. http://labs.strava.com/heatmap/#15/-76.83160/38.64195/gray/bike

The “heat” on the most popular trail out there just ends for no reason. There is nothing there at all to cause this, no place to turn, or even a good place to turn around. Plus the primary parking area is within the area shown with little riding heat.
[ATTACH=CONFIG]10434[/ATTACH]

[Raymo853]

Another, and possibly understandable in light of National Security demands, are the roads on Andrews Airforce Base. There is a big drop off in the northern sections, and there are no gates at the locations of the changes.
[ATTACH=CONFIG]10435[/ATTACH]

http://labs.strava.com/heatmap/#15/-76.87726/38.82246/gray/bike

[Tim Kelley]

Here’s another one: http://labs.strava.com/heatmap/#12/-76.09955/37.01174/gray/bike

Who is riding on the Bay Bridge? And why is the northern part more ridden than the southern part? And how do you get GPS signal in a tunnel?

[Raymo853]

@Tim Kelley 131814 wrote:

Here’s another one: http://labs.strava.com/heatmap/#12/-76.09955/37.01174/gray/bike

And how do you get GPS signal in a tunnel?

Smart phones use more than GPS to locate and track you. Cell phone makers, service providers and consumer mapping products (Google Maps, IMap, Waze) actually prefer to use the other sources of your location (WiFi maps, triangulation from cell towers, their own user location heatmaps,…) as it is quicker, less math intensive, and uses less energy while hopefully offering them an edge on knowing where you are over other companies.

[Tim Kelley]

@Raymo853 131821 wrote:

Smart phones use more than GPS to locate and track you. Cell phone makers, service providers and consumer mapping products (Google Maps, IMap, Waze) actually prefer to use the other sources of your location (WiFi maps, triangulation from cell towers, their own user location heatmaps,…) as it is quicker, less math intensive, and uses less energy while hopefully offering them an edge on knowing where you are over other companies.

Yes. And can you confirm that Strava uses wifi maps, triangulation from cell towers, and their own user location heatmaps to record tracks?

[mstone]

@Raymo853 131821 wrote:

Smart phones use more than GPS to locate and track you. Cell phone makers, service providers and consumer mapping products (Google Maps, IMap, Waze) actually prefer to use the other sources of your location (WiFi maps, triangulation from cell towers, their own user location heatmaps,…) as it is quicker, less math intensive, and uses less energy while hopefully offering them an edge on knowing where you are over other companies.

Apps prefer coarse location services for things like weather forecasts or ad targeting, because it consumes much less power than gps. Wifi tends to be more accurate than cell tower location, but neither is anywhere near as accurate as gps. (Cell location is typically accurate to miles, wifi hundreds of feet, gps to tens of feet.) I don’t know of anybody who attempts location plotting (e.g., strava) via coarse location.

[worktheweb]

I would guess that the majority of “highway riding” was people leaving Strava on after finishing a ride and then driving from the end-point, or using it to track car rides (not sure why you would want to do that). I’ve seen this from many users who get a KOM that is then quickly flagged.

As for going through the tunnel, Strava seems to connect the points between GPS reads. I’ve seen people with really odd straight-line tracks when their phones are not getting good GPS locks. The ride through the tunnel is probably Strava doing a straight-line connection between the last track before and the first track after the tunnel.

[Tim Kelley]

@worktheweb 131832 wrote:

The ride through the tunnel is probably Strava doing a straight-line connection between the last track before and the first track after the tunnel.

Excellent thinking–you’re probably right on that. And then why does the line fade out so much in the southern half?

[Author]

Replies