I’ve been doing lots of GPS legwork in preparation for a piece of place-specific work I’m developing for a section of Digbeth.

Basically this means walking a fixed route many times whilst carrying a GPS device so I can then look back at the trace of where it thought I went and get a feel for how much the buildings deflect the GPS signal. The things I’m particularly interested in (sorry – can’t be too specific at this stage!) are extremely liable to do this, so let’s just say I’m spending a lot of time walking around Digbeth at the moment!

Each walk takes approximately an hour to complete: I walk up and down particular roads, always staying on the right hand side and walking in the middle of the pavement (or as close to it as I can get seeing as how the pavements are often substitute carparks in Digbeth).

A few days ago I did a nice walk in the sunshine, but it was sullied somewhat when I got home and processed the traces – only the first 400m or so of my walk had been recorded! After that I decided to take both of my devices out with me, so I had a safety net if one of them didn’t record properly.

Yesterday I walked with a device in each hand:

iPAQs and gum

Check out those gum constellations! nice.

As I walked, I noticed that the device in my right hand was consistently giving me more erratic readings for my position than the one in my left hand. At one point I left them both next to each other on a wall for a few minutes so that when I could tell if it was my body disrupting the signal, or if it was because the device in my right hand was closer to the buildings I was walking past.

side by side


As you can see in the above screen-grab, there’s still a lot of discrepancy.

This got me thinking: by taking two devices out with me, I now had two traces that could be synchronised because the data that’s logged also includes time.

With a bit of fiddling around in notepad and calc I was able to merge the two sets of data and combine it with some mark-up so that that rather than having a line per device that joined position at time1 with position at time2, I could generate a series of lines that joined position from device1 and time1 with position with device2 at time1. Approximately.

Here’s a screen grab:


Most of the lines really aren’t that short though and Digbeth is criss-crossed with a mesh of yellow. The best attempt so far at describing it comes from Tom Maillioux:

Like you reprogrammed your GPS to make not just a point, but a line from everywhere you were, showing all the places where you COULD have been.

I love how this has turned the data-collection into a drawing project and how it’s resonating with aspects of psychogeography, belief in technology and the whole sort of indecisiveness of Digbeth.

Now all I need is a way of extracting those lines from the .kml file so I can print them out big.