We’re celebrating seven years of OS OpenData, and its success is down to the people and businesses using the products. We are always interested in hearing how open data is being used, so please keep sharing your examples with us. One business who we have spotted using our data regularly over the years is Parallel. We asked Ashley Clough, founder of Parallel to explain how OS OpenData has benefitted them.
Parallel has evolved to specialise in data-visualisation and mapping, particularly for healthcare data in and around the NHS. We started to use OS OpenData when we became frustrated by the styling of available basemaps for website applications. We needed a set of maps that were optimised for the presentation of data overlays; as icons for point locations and polygons for area indications. We needed to control what was visible on the map at every zoom level and crucially we needed to ensure that the level of detail was consistent across the entirety of Great Britain. We investigated using open source map data but we couldn’t rely on the consistency of data within urban locations, and particularly in more rural locations. As the maps are used within the NHS we needed to ensure that everywhere had the same quality of data; OS OpenData was, and we believe still is, the most consistent for our purpose.
We originally used OS VectorMap District alongside some vector data, but when the initial version of OS Open Map-Local was released in 2015 we quickly reworked our basemaps to use it as the primary source of data. We have three stylings of the core data: a primary map with a white background and limited colours; an inverted version with a dark background; and a mid-grey version that allows for some hill shading to show basic 3D landscape modelling.
We also have to confess to using some additional OpenStreetMap data in some of our basemaps. As good as OS Open Map-Local is, it runs out of detail at higher zoom levels when we need to display more detail with regard to land usage, public footpaths and fence lines. We use additional data sparingly but it’s still important to us to literally fill in some of the gaps.
Creating maps with OS OpenData
We use Mapbox to deliver the basemaps, some of the data overlays and the API for presentation and control of the maps on a web page. We’re committed to using open source solutions wherever possible. Technically, we load the raw OS data into a series of PostGIS databases – converting the data from the standard OS projection to WGS Web Mercator projection. We then filter and combine the data into a series of source ‘tilesets’ that are uploaded and then layered and styled in Mapbox Studio.
Typically, in our ‘roads’ tileset for instance, we will define which set of roads are included in the data at each zoom level (motorways at zoom 6, primary roads at zoom 7, A roads at zoom 8 etc) and then how each road is styled and named at each level (simple single lines at low zoom ranges, stroked thicker lines at higher zoom levels). The entire set is then composited into a single Mapbox ‘style’ that can then be called from Mapbox as a set of tiles for use in a web page or web application. One benefit of this ‘layered’ approach is that it allows us to effectively ‘sandwich’ our data layers between the base layers of the map and the layer that carries the location names – effectively ensuring the names are always visible above the data.
Projects using OS OpenData
Our key application is SHAPE (Strategic Health Asset Planning & Evaluation) for Public Health England. Core SHAPE is a free-to-use application for the NHS and Local Authorities and has been in constant development for the past 12 years. More recently a series of atlases have been developed that provide a range of options for visualising the location and details of the healthcare facilities in England; looking at specific datasets relating to specialised clinical commissioning, dementia and public health; and presenting Local Authority care provision and demographics.
The newly released SHAPE Place atlas brings together a range of mapping capabilities centred around the location of the healthcare estate, the accessibility of the sites via a range of transport options and the demographics and clinical needs of the local population, including details of each GP practice’s registered patient population. Key to the presentation of the data is the use of our OS OpenData base maps.
Alongside the maps we make for our clients, are a range of maps that we are producing to explore the possibilities of next generation 3D vector mapping. These are enabling us to render demographic data: deprivation and population projection as ‘stacks’ representing counts of people within LSOAs (Lower Super Output Areas) across England and Wales; to display very large datasets (like the 1.7 million Code-Point Open postcode locations); and to visualise 3D building heights and functional site classifications in key cities (images at top of blog and below).
Why use OS OpenData?
We’ve benefitted directly from the increasing level of detail that has been included in each iteration of the OS OpenData and can see the changes that are being made as OS migrate their resources to satisfy both paper-based presentations and live digital environments. Of course, I think we’ll always want more detail, road widths and data that identifies vertical relationships when roads and railways cross would help take the data from being purely representational to much more accurate at a local level. The building height data that is available in other OS product sets would also be welcome as we’d like to be able to use our 3D maps for the whole of the country, not just those areas where we have used other sources of data to enable 3D rendering of buildings.
Having said that, the two clear benefits for us with OS OpenData are the consistency of the data and its easy, free licensing. We can download and process the data to create a single data source that we can then style and use appropriately as demanded by the needs of the map’s users. Integrating the data with other open data sources is critical; we constantly use sources like ONS and NHS Digital, amongst others, for the supply of datasets that we then visualise over our basemaps. OS OpenData is very probably our most important dataset.
See more of Parallel’s work: https://parallel.co.uk
Get started with OS OpenData: https://www.ordnancesurvey.co.uk/business-and-government/products/opendata-products.html