Limits and considerations of Strava’s data In this guide, we will outline some of the complementary methods and tools to analyze the unusual data you may find on the Strava heatmap, and a consideration of some of the limits of the recorded activity. However, the map alone is sometimes inadequate to provide useful analysis. This is where I politely remind that it is sitting on a ton of data that most intelligence entities would literally kill to acquire. US Bases are clearly identifiable and mappable /rBgGnOzasqĪs The Verge wrote, you can already find most of these military sites with satellite imagery, but the heatmap reveals human activity, such as “ how people are moving along those areas, and how frequently, a potential security threat to personnel.” As Jeffrey Lewis noted, the data held by Strava, both in its heatmap and elsewhere in its databases, is invaluable. … It looks very pretty, but not amazing for Op-Sec. 13 trillion GPS points from their users (turning off data sharing is an option).
Nathan Ruser was the first person to notice the potential use of the Strava heatmap to identity military sites. Here's the code, if you'd like to download it and try running & modifying it. It's a nice little example that shows how to add a lot of things to a map. And I use Proc Sql to grab the maximum distance (117.7 miles) and stuff it into a macro variable, so I can use it in the title. The landmark labels have html drill down links to the google map zoomed in to that location. I have labeled certain cities and landmarks based on their lat/long values, with the location of the label determined by an x/y offset.
The lat/long grid lines are created programmatically in a data step, and annotated on the map. There are a few other things I added that are also worth mentioning. Below is a snapshot of the map - click it to see the interactive version with html hover-text:
I added some extra variables to the dataset so it could be used to annotate markers on the map, and added html hover-text so you can the timestamp and distance for each marker. I used the lag() function in combination with the geodist() function to calculate the distance between each successive data location reading, and I used a data step to calculate the running total miles. OUT=anno_path ( rename= (var1=date_time var4=lat var5=long ) ) PROC IMPORT DATAFILE= "cuba_crossing_spot_data.csv" Here's how the csv data was structured:Īnd here is the code I used to import the data into SAS: So I had my buddy Mark download the data in csv format, so I could do my own analysis. But I wanted something a little different. The SPOT map was very useful for interactive/instantaneous tracking - it lets you pan & zoom, and view the map or a satellite image, and you can view the data points 50 at a time. This device tracked our position, and relayed the coordinates directly to a satellite every ~10 minutes, so our friends, family, and fans could monitor our progress on SPOT's Web page. Hopefully this will inspire and help teach you to create your own custom maps!īefore we get into the nitty gritty data analysis, here are a few pictures I took in Cuba.įor our trip (kayaking the 100+ miles from Cuba to the US), we used a special tracking device called a SPOT (on loan from my good buddy Mark - thanks Mark!). This blog post shows how I created a custom SAS map of the tracking data for my recent trip to Cuba. These days many devices (such as smart phone apps, Fitbits, Apple watches, dog tracking collars, car gps, hiking gps, teen/car trackers, etc) can track your location, and provide you with standard/canned ways to analyze the data.