OSSpatialResearch

This is the repository for Open Addresses' "Spatial Research" Python scripts, data files and SQL used during our evaluation of Ordnance Survey Open Map and Open Roads , part of the solution we deployed to ascertain their potential to improve the accuracy of our services.

System Requirements

Recommended:

Min 2 core processor at 3Ghz (4 core recommended)
min 8GB RAM (16GB recommended)
60GB free disk space
Linux or Windows with Python 2.7
MySQL, MariaDB or compatable database with full spatial query support

Dependencies

The Python scripts require the external PyShp shapefile Python library. Consult the installation guide for your OS for instructions.

They also require the common bulkinsert.py and extract_shape.py scripts included in this repository.

Running

Clone this repo:

git clone git@github.com:OpenAddressesUK/OSSpatialResearch.git

Create a database

Create a MySQL, or compatible RDBMS, database with spatial support and default character set utf8-unicode

Create a user on the database with admin rights to create, delete and truncate tables.

Amend the Python scripts to add the server, user, password and database name parameters.

Populate the database

Download the latest version of OS Open Map Local and OS Open Roads from the Ordnance Survey Open Data Portal

Unzip both files

To load OS Open Map, run the command

python load_OpenMap.py *directory*

If directory is ommitted, the script will run in the current directory.

The script will recurse the directory structure and aggregate the Os grid tiles, creating one table per shapefle type in Open Map with the Geometry field appended.

Fix missing data

Run the RoadFx.sql script on the database after it is loaded.

This will create a look up table from which the missing components in Open Roads can be added.

Test the database

To test the database, run the Python script getAddrFromOS.py:

python getAddrFromOS.py *postcode* *maxdist*

Where:

*postcode* is the postcode to be search without any spaces
*maxdist* is the maximum distance from the postcode centroid to search for address components

Example

python getAddrFromOS.py EC2A4JE 100

Yields output:

    Street: Clifton Street
    Settlement: 
    Postcode: EC2A 4JE
    County: Greater London
    Distance: 30.4
    
    Street: Holywell Row
    Settlement: 
    Postcode: EC2A 4JE
    County: Greater London
    Distance: 30.4059204761
    
    Street: Worship Street
    Settlement: 
    Postcode: EC2A 4JE
    County: Greater London
    Distance: 36.5766804691
    
    Street: Paul Street
    Settlement: 
    Postcode: EC2A 4JE
    County: Greater London
    Distance: 58.6853074283
    
    Street: Bonhill Street
    Settlement: 
    Postcode: EC2A 4JE
    County: Greater London
    Distance: 63.4154555294
    
    Street: Wilson Street
    Settlement: 
    Postcode: EC2A 4JE
    County: Greater London
    Distance: 70.5543761931
    
    Street: Dysart Street
    Settlement: 
    Postcode: EC2A 4JE
    County: Greater London
    Distance: 82.023204316
    
    Street: Finsbury Market
    Settlement: 
    Postcode: EC2A 4JE
    County: Greater London
    Distance: 88.1981859224

Licence

This code is open source under the MIT license. See the LICENSE.md file for full details.