County Confusion

Hello,

These are postal locations rather than geographical.

An example....

I live in Henley-in-Arden in Warwickshire which is 5 miles into warwickshire. However I have a Birmingham postcode B95 that means my postal address is West Midlands. The next village (less than half a mile) has a Coventry postcode and Warwickshire postal address.

Phil,
I did this a while back, I can send you a script that will sort the POI's into Type (LR,CR,PH etc etc) and county (based on the geographical regions based on the GPS waypoints around each county). Currently they are stored in a SQLliteDB but it's trivial to push it to mysql instead. Out of 3000 POI it manages to parse and 'correctly' decodes the type from the textual descriptions - the ones it fails on are listed separately and are due to the earlier POI being in a different format.

Its quite useful to use this data as I have now my own filtered set of POI into whatever format I want (I ignore LU for example).

More than happy to help out.

James
 
There must still be some County Confusion, I think, as,
though I nearly died in 1998, I lived on ,
and it wasn't until 2004 that I was sued by one of the Welsh Counties that, according to BigPeetee, " died in '96".:(:help::sad::wave::cool:

It would have been a strange scene in court if both parties had actually died in the preceding decade. :dnd:

sean rua.
 
Phil,
I did this a while back, I can send you a script that will sort the POI's into Type (LR,CR,PH etc etc) and county (based on the geographical regions based on the GPS waypoints around each county). Currently they are stored in a SQLliteDB but it's trivial to push it to mysql instead. Out of 3000 POI it manages to parse and 'correctly' decodes the type from the textual descriptions - the ones it fails on are listed separately and are due to the earlier POI being in a different format.

Its quite useful to use this data as I have now my own filtered set of POI into whatever format I want (I ignore LU for example).

More than happy to help out.

James

I have already done it, but thanks for the offer.
I used google maps to get the data but there were a few odd results.
 
I'm curious, how did you do it with google maps? My approach was to get county border data and then use maths to check if each point resided inside the borders. I thought about using address lookups but didnt think that would be as accurate.
 
Hi I am ignorant of programming so this may be a daft suggestion but if you go to the met office they have simplified by doing basic areas of the country and on the mobile site at least if you go five day local it asks , giving maybe a dozen options for the whole country, then it asks for first three letters of place name then gives you matching towns and you choose yours.
I just wondered if using their areas would help simplify things
 
I'm curious, how did you do it with google maps? My approach was to get county border data and then use maths to check if each point resided inside the borders. I thought about using address lookups but didnt think that would be as accurate.

I used the google maps api. This returns either a xml or json result containing data about the latitude and longitude provided.
I then parse the data to get; route, locality, district, post town, county, administrative country (England, Scotland, Wales, etc), closest postcode and post code prefix. Then insert the data into a MySQL database.

I then fetched the DISTINCT counties from the database and separated them by country.

There were some results that were not counties so I found the correct county and updated the database.

If you want more info let me know.
 
Cool, i didn't realise the google API gave so much data out, that is pretty handy. Another thing my tool does is grab the streetview locations (where available) so you can get an idea of the area (i grab 4 images for roughly N/S/E/W).
I have also got OpenStreetmap db installed locally and I have got the workings of a tool that can get local amenities for each POI, although this is not quite finished yet. For example some of the open streetmap data has markers for water taps, public loo's etc.

Also played with rendering custom map tiles to show just the stuff thats interesting to us as motorhome users, but that has proven extremely time consuming and something I have put on hold for now.
 
One of the reasons I avoided looking too much into the google maps was their License agreement for the data is not that friendly iirc, hence I was looking more at the open street map stuff, just a thought to keep in mind if your project will turn commercial in any way.
 
The reason I want that data is so I can filter the pois by 'region' when browsing or planning where to go. For example, soon I will need to goto Birmingham so I want to see all the options without needing any prior knowledge about postcodes etc.

For the GPS files for sat navs this may not be so useful.
 
I don't need the postcodes for my approach to work. Postcodes are not free data either, there is open postcodes websites but they are not complete or accurate (at least for where I live) so legally I would require a license from the post office which is extremely expensive, or rely on user submissions.

Regardless, simple bit of Maths let's me identify which county a given gps point lies in, and I'm not bothering to normalize the results into other tables, not much point for the relatively small data set I have. I can just query the pois from a flat table.

This has all been done already for myself, I was just curious on Phils approach to figuring out where the points where (county, postcode etc).

Still I have a nice saved view showing me all pub stopovers in whatever region I need :wacko:
 

Users who viewed this discussion (Total:0)

Back
Top