This is Bryn, a 7-month-old, infinitely wild, working cocker spaniel, and for the past month, I have been mildly obsessed with tracking his movements.
I spend most of my days creating data-visualisations based on the movement of people and vehicles so it was only natural that as soon as Bryn came into the family I was googling GPS devices to attach to him. Seven months on and with some good puppy training instilled I was ready to track him.
Being a working cocker he is quite loyal in his movements, he will stick quite close to me or my wife and constantly seeks reassurance before venturing off. Whilst this is interesting in its own right it doesn’t make for a particularly interesting visualisation. However, Bryn has recently discovered a neighbouring cornfield and with a nod from the local farmer, he began exploring.
For the first few walks, he would disappear for a couple of minutes at a time into the maize which prompted me to get a tracker for him so I could see where he was exploring. At the time I only had a Garmin Forerunner 235 watch which I was a bit nervous about attaching to him, something I would later learn to regret! Needless to say, I fixed it onto his harness and off he went, the below is the first foray into #plotmypaws!
So, admittedly it is flawed, my track is a recorded track from a previous day so doesn’t really play correctly to where I was in relation to Bryn. However, it is interesting to see Bryn’s patterns deviating in the cornfield (the easterly field), he becomes braver and his movements are far more erratic.
What’s the next most logical evolution of #plotmypaws? Mapping his movements over 1 week!
Bryn gets walked twice a day, early morning and with the heat at the moment, late at night. On each walk, I attached my Garmin to him and started tracking him. As my Garmin is linked up to my Strava account it has been quite funny watching him get ‘Personal Bests’ for fastest mile etc, the Garmin even tried to record his heart rate for a ‘Suffer Score’ but not quite sure how it’s tracking that!
After every walk I download the GPX from Strava, bring it into QGIS to anonymise my origin (something I forgot to do at first!) and give each track a unique ID. After I export the GPX points to .csv I’m able to plug the data into our in-house ‘Design Suite’ to create the motion trails.
The process went smoothly until Bryn appeared without my watch on his harness! My pretty pricey Garmin watch was now sitting somewhere in the below…
Retracing my steps seemed impossible seen as Bryn had no real pattern to where he could go, but there was one glimmer of hope, my phone is linked via Bluetooth to my watch so when they are in close proximity they sync up! I ended up walking very slowly around that field for a good hour until it synced and then I was able to find it — thank god for connectivity!
Obviously I decided to purchase a cheaper tracker, and ended up with the Renkforce GT-730FL-S GPS Data Logger.
Don’t be fooled by the user reviews, most are complaining about the packaged software and not the quality of the device. Generally, it’s pretty good, my Garmin watch sampled at around 5-second intervals, the Renkforce can sample at 1 sec intervals. Although sampling at such a high rate can expose its GPS inaccuracy a little more than the Garmin. It has 2.5m CEP accuracy which is ok, especially for what I am doing, and better than the Forerunner which is pretty inaccurate (http://fellrnr.com/wiki/Garmin_235#GPS_Accuracy). It also acquires satellites almost instantaneously and, until now, I haven’t had any noticeable dropouts.
So 1 week and 14 walks later I was able to produce this:
Extrapolating my walk distance puts me around 35km over 14 walks, Bryn almost doubles this at 67km. It’s not only beautiful to watch but is also pretty interesting to see that he has favourite areas he likes to explore, the heat of the trails shows notable patterns in the North and North-West corners of the maize field. You can also see linear patterns when he deviates from the path — this is him running down the rows of corn.
Graphically I added some contextual data into the scene; buildings and trees, and also a new logo and strapline ‘Plotmypaws’ which sounds more catchy than ‘Dog Viz’!
Where next for Plotmypaws? Well, I now have two functioning GPS trackers which means I can get down to the detail of how far Bryn is deviating from me at any one point. This in itself is interesting, but what if we compare how his behaviour changes when I or my wife walk him?
Now, the below is fundamentally flawed in that I am just using one sample for each so it’s not fair to derive any meaningful relationships from the visualisation.
However, it forms the basis of what could be a really great visualisation if I (and my wife!) persist in dual-recording our walks every day. That said, what if there are fundamental differences between the time of day we walk him? My wife can only walk him in the evenings so what if he is generally more active in the evening than in the mornings? There are obviously many factors that we would need to take into consideration but for the time being, I am just treating it as a bit of fun.
The visualisation should be straightforward, I designed a little viz element that bridges the two tracks with a dynamic circular buffer and connecting bridge. It shows how far Bryn deviates from both me and my wife at any point during the walk. This is then transposed into a graph so you can see how that looks in terms of time vs distance. Whilst he deviates a lot more during the ‘corn-field’ part of the walk he always doubles back and returns to each one of us, which is typical spaniel behaviour. Whilst he seems to run off further from my wife the max distance he ever goes from either of us is pretty much identical!
I may have spoken too soon, yesterday morning Bryn decided he didn’t want to respond to my whistle and calls and about halfway through the walk we lost each other. It’s tricky to see in the above GIF but the only reason I return is that my wife called me telling me Bryn had come home! It looks like we just missed each other at one point. Fortunately, there are no main roads but it was enough to make me want to put some sort of beacon on him!
I’ve loved tracking Bryn over the past month and whilst I think I am going to call an end to the Plotmypaws series (for a little while anyway). I do think there is more analysis/visualisation that can be done here…
- Visualising the difference in Bryn’s tracks pre and post corn harvest
- Visualising the difference in Bryn’s tracks in the evening and morning walks
- Tracking different dog breeds to see how they deviate from their owners.
- Bryn’s dance over a month/year.
- Movement patterns of my 3 house rabbits 🙂
Bye for now!