4 ms·

I started a company called locatable (www.locatable.com) with a couple of mates some years back. The search engine did the same thing in the UK. The site is now

I started a company called locatable (www.locatable.com) with a couple of mates some years back. The search engine did the same thing in the UK. The site is now offline, but there's some screenshots here https://data.gov.uk/apps/locatable-0 https://data.gov.uk/apps/locatable-0 By far the most time consuming aspect of running the algorithm was computing the polygons to plot on the gmap. The actual search is fairly straightforward, even when using roads instead of the relatively simple public transit graph to compute the isochrones.

kawsper 11y ago
Nice - I wanted something like that the other day. Hopefully with open government data we would some day be able to solve it. I want to specify "I want to live close to a forest, a skatepark, and with an airport 1.5 hours from me", and then it should give me nearby areas that fit the criteria. Data-collection, cleaning data, and tagging it up is hard though :(
- chillydawg 11y ago
  Actually, the data cleaning isnt so bad. Open Street Map is what you want. I wrote a piece of software for locatable that parsed a given chunk of OSM (in this case, all of the UK mainland) and pull out stuff like hospitals, transit stations, land usage types and so on. I had two models: a zone and a poi. Zones were an area, eg a park or a forest or an industrial park that you could specify as an include/exclude + buffer distance. Eg: At least 10mi from a forest. Poi is point of interest and had the same search semantics: put me 1mi from a GP surgery or a supermarket. Or put me at least 1mi from a primary school because..er...mistakes were made...
poooogles 11y ago
Oh this was amazing, I used it a couple of times when I first moved to London to figure out where I needed to live. It helped out massively compared to the complete guesses I was going on.
- chillydawg 11y ago
  Glad to help :)

kawsper 11y ago

Nice - I wanted something like that the other day. Hopefully with open government data we would some day be able to solve it. I want to specify "I want to live close to a forest, a skatepark, and with an airport 1.5 hours from me", and then it should give me nearby areas that fit the criteria. Data-collection, cleaning data, and tagging it up is hard though :(

chillydawg 11y ago
Actually, the data cleaning isnt so bad. Open Street Map is what you want. I wrote a piece of software for locatable that parsed a given chunk of OSM (in this case, all of the UK mainland) and pull out stuff like hospitals, transit stations, land usage types and so on. I had two models: a zone and a poi. Zones were an area, eg a park or a forest or an industrial park that you could specify as an include/exclude + buffer distance. Eg: At least 10mi from a forest. Poi is point of interest and had the same search semantics: put me 1mi from a GP surgery or a supermarket. Or put me at least 1mi from a primary school because..er...mistakes were made...

poooogles 11y ago

Oh this was amazing, I used it a couple of times when I first moved to London to figure out where I needed to live. It helped out massively compared to the complete guesses I was going on.

chillydawg 11y ago
Glad to help :)