Article: "You can map Britain and Ireland's population centres, using only the location of pubs"

[Kim Foale]

Really interesting data vis from a lateral thinking perspective! Wonder what else could be done in this kinda indirect way?

http://www.citymetric.com/skylines/you-can-map-britain-and-irelands-population-centres-using-only-location-pubs-1286

Was reminded of a thing I read during my PhD research. Some city in Canada did 100,000s of acoustic measurements over an entire city over decades with dozens of RAs to see how much the background noise levels were increasing. A soundscape researcher went to an ambulance museum and measured how loud the sirens were. Came to same thing: 1.5dB/decade (IIRC), but going back another 50 years. 

Food for thought!

Kim

[Robin Gower]

As far as I can tell, pretty much everything is correlated with population which means you can kind of infer anything from anything else (at least when it comes to the spatial distribution of human activity). Indeed some physicists even derived a model for a city based upon population alone: http://www.radiolab.org/story/91732-cities/ (can you believe even walking speed is related?!).

I suspect the background noise rose in proportion to traffic levels which grew as the city grew with siren volumes rising to compensate.

This can be a challenge in econometrics because you can find false-positive correlations everywhere - everything grows together. That's why you typically need to denominate time series (typically by population) or look at rates of change.

I would have thought the pub visualisation could be much more insightful if it displayed pub density instead (e.g. pubs per capita in 15 minute walking radius)... indeed I think BBC R4's More or Less looked at that a few years ago but all I can find now is this post explaining why that kind analysis itself is plagued with problems (which population? etc): http://news.bbc.co.uk/1/hi/magazine/7370986.stm.

It would be interesting to know if ambulance sirens got quieter in London after the congestion charge cut back traffic in town...

--

Robin Gower    |    +44 791 255 3187    |    infonomics.ltd.uk

3DE8 5FEF 8662 FE92 3119  2E6F F59A DD96 5A90 A1B6

How to improve the quality of your data - lessons learned from analysis, explained with the science of information entropy

--

---
You received this message because you are subscribed to the Google Groups "Open Data Manchester" group.
To unsubscribe from this group and stop receiving emails from it, send an email to opendatamanches...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Julian Tait]

I used to work for a large well known pub chain and they used a lot of data to make decisions on where there pubs would be allocated.

They calculated return on investment based upon demography and locations where found based on yield within an area. A lot of chains do this, that is why there is a tendency for them to clump together.

So the data creates the pub from which we infer where the population is :)

J