CSN coverage map (as of July 2017)

CSN coverage map (as of July 2017)  

  By: ijl20 on Sept. 29, 2017, 3:52 p.m.

Here's the area covered by our current Cambridge Sensor Network CSN heatmap

Re: CSN coverage map (as of July 2017)  

  By: straytaoist on Oct. 8, 2017, 11:33 a.m.

Is there any way you can host this map on the site so it can be played with? I'm pondering my work cycle commute, and if I can pass through hot spots, rather than dips. (And as this is an image, and even though I know the area, it is awkward to play with.)

Re: CSN coverage map (as of July 2017)  

  By: ijl20 on Oct. 20, 2017, 1:25 p.m.

Is there any way you can host this map on the site so it can be played with?

The map has been implemented as a web page which may in due course make its way onto the SmartCambridge platform, with the idea that the Cambridge Sensor Network coverage would be derived from sensor signals received. The actual implementation uses a 'hard-coded' dataset that was created by a dozen volunteers who drove, cycled and canoed their way around Cambridge in May 2017 with a Adeunis Test Lorawan sensor so we could get the baseline data.

The usual method used in wireless networks is derived from the antenna positions and a topographical map of the area - these look really professional and detailed but I don't trust the result. Painting the heatmap from actual sensor data is better IMHO so that's what we want to do.

In the meantime, the coverage heatmap prototype webpage version 'minus 100' is on my University webspace here: heatmap webpage

You can specify a date range for the data used with 'date_from' and 'date_to' querystring parameters, e.g.


You can unpick the actual Adeunis Test data from the heatmap page, or if it helps the actual datapoints collected are displayed here: http://people.ds.cam.ac.uk/ijl20/smartcambridge/csn_survey/

The code takes the 'scatterplot' of the actual data, normalises it to a regular grid, and then uses a Leaflet heatmap library to blur that into a colour-coded heatmap. That kind of processing is routine to Geographers but they always use desktop tools such as 'R Studio' and we wanted a solution that can be rendered live in a web page. Hence the inelegant bucket of Javascript you can see in the page source. By the time it goes on SmartCambridge.org it will have been re-written and use an as-yet unimplemented API to get the data.

It's not perfect but it's better than the '3km circles drawn around the antennas' that we had before.

 Last edited by: ijl20 on Oct. 20, 2017, 1:32 p.m., edited 6 times in total.
Reason: added links to web pages