UK 3G data coverage in national parks and protected areas

by tokumine

Understanding 3G data coverage in remote areas is pretty important for environmentalist geeks, as typically our mobile apps will target use outside of urban, habited areas. To ensure our apps work, we have to know how much we can rely on the internet and cloud based data processing and storage services through mobile 3G data services. Ideally we’d have some numbers too.

Last Thursday, I presented some back-of-a-napkin GIS work on 3G data coverage in UK national parks and protected areas as part of the Symposium on Mobile & Conservation at the Oxford University Biodiversity Institute. I’ve written up the methods below, but the take home message is that ~18% of UK protected areas have 3G data coverage, assuming Vodaphone has a similar pattern of network distribution across the UK to other mobile operators.

As far as I can tell no UK mobile telecoms company freely releases their 3G coverage data in an easy to analyze GIS format. I had to do some fairly hairy data-munging to get this data, therefore the results should at best be considered a starting point for further work with official released data.

The best official 3G coverage data I found was a PDF from Ofcom in 2009 (useless for analysis). Of the current data provided by mobile companies, non provide a queryable interface and in general provide maps which show a really optimistic mix of data and non-data service coverage. The maps are pretty tho.

3G coverage optimism

Optimistic mobile coverage

Despite the worlds worst interface, the vodaphone coverage checker does allow you to choose a map of only 3G data coverage, and is updated weekly (data used was updated 11/01/11)!

Voda 3G

Voda 3G

After a quick root around in the protocols my browser was firing to update the map using the Charles proxy, I was able to get the Vodaphone tile server to return a single tile of 3G coverage. From here, it was a simple step to write a quick script to scrape off all the tiles for the UK and assemble them in an image editor.

UK Voda tile

UK Voda tile

I then isolated the color range for the 3G coverage, and (very roughly – note NI!) georectified it in ArcGIS for use in a spatial analysis (with some help from @craigmmills)

Vodaphone UK 3G data coverage

Vodaphone UK 3G data coverage

Finally, intersecting with spatial boundary information from all 8944 UK protected areas from the WDPA yields a total 3G Vodaphone coverage of only 18% of UK protected areas, skewed to the south.

UK protected areas

UK protected areas

UK protected area Voda 3G coverage

UK protected area Voda 3G coverage (~18%)

/I’m pretty sure this is anecdotally what we’ve all experienced anyway: it is highly likely you will currently experience no or very limited 3G/data connectivity in UK Protected Areas, especially in wales or the north. I guess there’s simply nothing in it for the mobile telcos to provide connectivity at the top of Snowdon.

On a positive note, Ofcom is trying to open up the 2G airwaves to data services. This is a really great direction that I hope all networks will support through implementation. Note that there are significant areas of the UK (again, mainly in protected areas), where there is no 2G coverage to start with. It will be interesting to rerun this analysis at a later date  and see how this situation has changed.

If the UK is an example of a pretty well connected country this situation is likely to be similar/worse in other countries.

If you’re thinking of making an outdoors environmental mobile app

  1. Data reception is distributed heterogeneously
  2. The server side of the the modern mobile app (cloud services, heavy duty data processing via hadoop etc), is something that your mobile environmental app will not be able to depend on.
  3. Mobile environmental apps should be native, not web based, support offline sync, and do any novel data processing on the client handset.
  4. Although HTML5 shows some promise with some offline capability, I’m fairly sure going with a HTML5 app in a protected area would still be a recipe for browser-refresh based disaster.

If you replicate or develop this idea further, please get in touch, I’d love to hear about it. If you are a mobile operator and would like to provide me with different 3G coverage data, I’d be more than happy to re-run the analysis for the UK or any other country!

Thanks again to Paul, Michelle and all at the school of Geography and the Environment for organising an energising and thought provoking day.

Advertisements