A map of Farnborough in the UK, covered with pins to show where planning applications have been made

mySociety built a responsive web tool to help members of the public search and explore planning applications across the county, using data from Hampshire’s data store.

Challenge

Every day, thousands of planning applications are submitted to local councils around the country by people applying to demolish a garage, erect a fence or convert a loft. More often than not these applications disappear into proprietary systems that, despite being publicly available, make it hard for members of the public to find out what’s going on in their area.

In summer 2014, we were invited by Hampshire County Council to participate in a tender for an interesting data-driven prototyping project. Hampshire CC wanted to find a partner who could take a standardised set of planning application data they were compiling for planning authorities across the county, and make it “more understandable, accessible and useful.” Ultimately, they wanted to help members of the public and businesses better understand what was happening within their communities.

Initial Concept

We responded to the brief with a concept focused on helping members of the public, but which could easily translate into commercial use cases. We proposed that Hampshire should eschew fancy data visualisations and dashboards in favour of something more simple but still desperately needed in the planning space: a map-based browsing and searching interface.

hampshire homepagehampshire results 1hampshire result

Our concept had a slight twist, however. We proposed to use a Bayesian classifier (a simple algorithmic way of determining, in this instance, whether something falls into one type or another based on the words used to describe it) to categorise the applications into common types that are normally of interest to your average home owner.

We wanted to make it really easy for people to see all of the applications for, for example, conservatories or loft conversions near where they live, and quickly get a sense for which ones were being approved or rejected.

We hypothesised that this could help in many different use cases: for homeowners researching what makes a successful application for an extension; for a business looking to relocate within the county looking at the likelihood of a change of use application being granted; or for campaigners looking at applications granted for, for example, new supermarkets or convenience stores.

Data Exploration

After being awarded the project, we started with an exploration of the data to see if the approach that we had wanted to take would be possible, given the reality of its structure and content.

The data source for this project was initially planning application data from Rushmoor Borough Council, provided through a data hub built by Swirrl. This data had been organised according to a schema developed by Hampshire and Rushmoor in collaboration with the LGA and Swirrl.

We wrote a small bit of Ruby code to help with this. In order to make our application work, we needed to get the data from the hub’s linked data format into a relational database. This makes it easy to search and query. Our code takes a list of applications on the hub as a starting point, and then traverses through all of the linked documents to build up a complete picture of each application, from the key dates through to the final decision. With this, we could extract the data in the format that we needed in order to build a search engine, run our classifier and calculate statistics all from one central place.

Need a little help with your digital offerings? Get in touch

Design

We already had some idea of the likely structure of the application that we had established in our concept work. We needed to check, however, that this was still sensible now that we were actually grappling with the details of the implementation with our clients Dan and Mark.

Our starting point for this was a workshop with them and another expert in the data from Rushmoor Borough Council – in which, among other things, we went through designs for the key pages in the application and prioritised the different potential bits of information and functionality.

hampshire-workshop

We used our learnings from this meeting to put together a set of wireframes for the application. Our designer took these wireframes and the visual concepts for the new Hampshire Hub, and created a visual identity for the prototype application which we had decided to call “Open Planning”.

Development

Our first and most important technical decision was to build the prototype application on top of an existing open source codebase – Planning Alerts Australia. While this constrained some of the technical choices we could make for the project, it gave us a certain amount of functionality for free and will help ensure that the project and our work on it have greater longevity by being part of something bigger.

Planning Alerts is a Ruby on Rails web application, which provides functionality to store planning applications from multiple authorities and alert interested parties about them. We added the ability to view them on a map, search them, and to automatically categorise them based on their descriptions.

All of this work has been completed in the open, on the code collaboration platform GitHub. A de facto standard in open-source software development, GitHub makes not only the code, but the history of development, bug tracking and proposed features available for anyone to view and contribute to.

And now …

The prototype application has been completed; we are waiting for permission from Ordnance Survey to use location data derived from one of their mapping products before it can be shared more widely.

Dan Cooper from Hampshire CC said of the project:

“Working with the people at mySociety was a pleasure from start to finish. From the inaugural project meeting to the sign off meeting they were always approachable, professional and adaptable. The agile approach to the project we took was perfect when considering the short time frame and set value. Their understanding of open source ways of working, planning data and the community really helped hone the output into its final guise and without which the project would not have been as successful as we consider it to be.”

Would you benefit from mySociety’s skills and experience? Drop us a line