Introduction

The English and Welsh Police forces national crime mapping website launched in October 2009. The work was led by the National Policing Improvement Agency (NPIA) on behalf of the forces, it is the first national crime mapping site for England and Wales and the first time, as far as we know, that the 43 forces have worked together on a public facing website.

The NPIA asked us to develop the site for them. The launch was fairly high profile and the site is attracting a large number of visitors and comments, positive and negative. There has been a lot of interest and lots of questions about how we tackled the project, so we’ve written this to help explain the background to the project, how we handled it and respond to some of the feedback from our point of view.

Much as it was tempting for us to get involved, we stayed away from the political angles that have steered some aspects of the website. We’ve always been on hand to provide recommendations, advice and our own sometimes strong opinions, but the site belongs to the police, and they have the final say.

This is going to be a fairly long article to try to cover all the angles but if you have any questions then please ask and we’ll do our best to answer them.

Who are we?

Rock Kitchen Harris is an advertising agency with a strong online team, we’re based in Leicester, England. We were established in 1982 and built our first website in 1996. We’ve been making all sorts of websites and web applications ever since.

Crime mapping online

Police forces from all over the world have been mapping crime for their own purposes for many years; identifying which crimes happen where, and when, allows them to target their resources effectively.

In the United States forces have been publishing crime data online for several years. Many of these provide point data (where crime is mapped to a precise location), and are updated daily or in real-time.

In the UK a few forces have had their own online systems for some time, however systems were developing independently and it was difficult for visitors to compare results between forces in any consistent way.

In mid 2008 the Metropolitan Police developed an online mapping system covering their force area and in June 2008 the Home Secretary announced that, by the end of 2008, every English and Welsh force would map crimes online following consistent guidelines.

Why were we involved?

We began working with our local force, Leicestershire, in 1999. We developed a content managed website for them using our own system and have worked with them ever since to expand and improve their site.

In 2005 we built the main force site for Nottinghamshire and again we continue to work with the team there to manage and develop this site.

We developed a system for both forces that allowed them to upload force-wide crime data by area that was published as basic graphs and tables. Separately, we had worked with these and other forces to map and patrol areas using Google Maps.

When the Home Secretary announced online crime maps for all forces, it was a natural progression for us to get involved - we had many years experience working with police forces, experience in publishing crime data, and experience in mapping.

Several forces chose to develop their online crime mapping systems either with their own in-house team or using an outside supplier, but by September 2008 the majority of forces were still considering their options. They decided to hold a joint meeting where four potential solutions were presented in one day, including ours. We presented an early prototype of  CrimeMapper, a system with a focus on simplicity and ease of use for the general public and we were delighted when 30 of the 43 forces chose it.

The last three months of 2008 became an period of intense pressure. There were technical issues, political pressures and serious project management challenges (we’ll explain some of these below) but just before Christmas 2008 the last of our 30 forces went live with CrimeMapper.

Following the launch the NPIA realised that the forces had achieved a great deal in a short space of time and saw that the maps were generally welcomed by the public. It was clear that the systems as they stood in early 2009 were going to be developed further but that it would be preferable for these to be consistent across the 43 forces as this would be both more cost effective and easier for the site visitor.

Scaling CrimeMapper to a national level

Our local system, used now by 31 forces, was well received by both the forces and site visitors. Everyone appreciated that there was room for improvement but it demonstrated that it was easy for forces to manage and the public facing site was user-friendly and accessible. There were significant cost benefits from so many forces using the same system and forces worked together in a way many had not done before.

In Spring 2009 the NPIA asked us to take the CrimeMapper forward as a full national system and we began work. The site needed to incorporate much of what had already proved successful but it now had to allow national comparisons, visitors had to be able to move seamlessly across force boundaries and it needed to be ready by the end of September 2009. We began to assess what we already had, and how we could scale it to meet the demand of a national system.

It’s a web application

For the NPIA and forces the fact that CrimeMapper is a web app was a huge benefit. No one had to install software locally, changes can be made instantly across all forces and as an added bonus we built it in a way that has proved so easy to use that no force required any formal training - everything was dealt with through an online support system and occasional phone calls or meetings.

Software

Whilst RKH’s web development team started out many years ago using Microsoft’s IIS, ASP and SQL Server technologies, we soon moved to open source alternatives. CrimeMapper uses our custom framework built on the traditional LAMP stack - Linux, Apache, MySQL and PHP.

Chicago-based 37signals had a strong influence on our application development techniques - we have a small team focused on keeping things simple, portable, lightweight and flexible. CrimeMapper is built in such a way that it can be easily maintained and further developed in future.

Hardware

When it came to hosting CrimeMapper, we didn’t have to look far. We have been hosting with Rackspace in London for many years, where they provide us with the very best support and hardware available. They gave us the most relevant load balanced hardware, along with the reassurance that we have access to highly experienced engineers to answer our queries and meet our hosting needs.

Project management

We faced some real challenges in managing the project in addition to the technical development:

• We were working with 43 independent forces serving different local communities - intelligent, passionate and committed to achieving the very best for their own force
• Whilst each force had one key contact, we regularly had to deal with people of varying skill levels and backgrounds - project managers, GIS experts, web developers, performance analysts and police officers
• Some of the larger forces have their own in-house web development teams with their own standards and skills
• Our main contact was the NPIA where there was a small dedicated team that we were in constant contact with
• The site understandably attracted a lot of political interest, we were involved at the edge of this but it was an ever present consideration for our contacts at the NPIA
• Anyone who has worked with the public sector will understand the procurement issues that have to be dealt with. The NPIA procurement team worked closely with us to resolve some difficult issues

To effectively manage the project across the whole of England and Wales, we used 37signals’ Basecamp, a web-based project management system. For day to day support, we used a simple support form, powered by WuFoo.

We were helped by having worked with so many of the forces before and hopefully gaining some respect and confidence from them. The deadline also kept everyone on track and forced us all to concentrate on the key issues and not get bogged down with trivia. Basecamp worked pretty well for most people, though with the numbers involved it was never going to suit everyone. What it did encourage was general discussions between forces with forces often helping each other by sharing their solutions and best practice.

What we’ve been criticised for

There have been a few recurring criticisms that we would like to address here. Some of these are linked and many of them we understand and accept. As we said earlier we realise the site can be improved and over time, and working with the NPIA, we hope it will be.

Polygons

Forces defined the areas to be mapped and also uploaded KML files to plot these. This allowed us to set the site up quickly but brings with it a couple of important issues:

1. The accuracy of polygons on the map. With mapping web applications there is a trade-off between highly accurate shapes and speed and performance. In some areas we have simplified the map polygons to a point where gaps are appearing - we’re working on this balance with individual forces and can usually resolve any severe gaps.
2. Difference in the quality of supplied data. Each force is independent, and so the quality of polygon data varies between forces due to differing GIS software and capacity and abilities within force. This can lead to issues where two force boundaries meet. Again, this is an issue we’re working with forces to resolve where there are problems.

Area consistency

To allow national comparison we required each force to map crime to 4 levels, the first being force. Again this allowed us to set up the site quickly and easily but there are a few problems.

Forces can vary significantly in size, the most striking example being the City of London which is the smallest force and is surrounded by the Metropolitan Police, the largest. It is impossible for all forces to have the same area structures so we had to use common sense and work as best we could. It isn’t perfect but we feel it works well.

Colours

This is the one issue that has attracted most criticism so far with some people telling us they have difficulty distinguishing the chosen shades. A number of requirements led to the colours that are currently being used:

1. Clearly it was important to use colours that users can easily distinguish on screen, with a reasonable monitor those chosen will work for the great majority
2. The colours should not cause a problem for those with colour blindness, using shades of a single colour met this need
3. Colours should work when printed on a mono printer. Had we used red/green for instance then these could be hard to distinguish when printed
4. Some colours were seen as emotive, making high crime areas red for instance could cause complaints from those who could say that the site my help to create no go areas

We will be reviewing the colours used in light of the feedback we receive.

Mapping provider

The choice of the provider of the underlying maps was a difficult and lengthy process.

A fundamental issue here is that forces frequently use Ordnance Survey (OS) systems to create their area boundaries, this is allowed under the licence agreement forces have with OS. The problem arises though when these areas are overlaid on an online map, some of the mapping providers terms are unacceptable to OS. For the local system the only solution was to create our own underlying maps working with an OS partner, ISYS Sytems. When the national system was developed an agreement was reached between NPIA and Bing (Microsoft) so we have worked with their maps and API.

Speed

The site provides a vast amount of information, statistical and graphical. There are many thousands of mathematical calculations involved and every time the user changes the map view or chooses another type of crime a lot happens behind the scenes. We have tried to reach a balance between providing lots of useful information in a clear way and giving visitors a site that works quickly. Apart from the performance issues on launch day we think we have reached a good balance, though from time to time there may be problems and some users won’t agree with us whatever we do.

Complexity vs Simplicity

We have been criticised for making the site too basic, we don’t agree  that this is a problem. Our primary aim throughout has been to build a site that is user friendly.

A year ago there were very few online crime maps in the UK, we now in a very short time have a national, consistent site. We took the view that it was essential to keep the site simple to use so we stripped out the complexity that it was very tempting to include, we’ll listen to all the feedback and, where it really makes sense, we will add more in time. Anther important consideration is what the forces can supply. Setting up this site was a major challenge for the forces and whilst some had lots of data and wanted to go further we had to produce something that was consistent and had to work with the forces that were least advanced, knowing we can change in future.

Launch Problems

For the first few hours there were problems with access to the site. We have to apologise for this. It was the aim of all involved to provide enough capacity to cope with a high level of interest, but we had to balance this with not wanting to over specify and waste taxpayers money. On the day interest was greater than any of us anticipated and the servers could not cope for a few hours. By the end of the first day it was all resolved and lessons have been learnt, it should not have happened though and we realise that.

Where We Go From Here

It’s far from perfect, we know. But in about 6 months 43 independent forces have come together and produced up to date statistics at force and local level, there is also lots of local information.

It makes sense at the moment to pause briefly and listen carefully to comments and see how the site is used and received. When you develop a site you become very close to it and it gets hard to stand back and see it with fresh eyes and objectively. We are really grateful for the feedback, much of which is constructive and positive.

If you’ve read the above you’ll understand that some of our main aims will be to make the site more consistent, easier to use, faster and to provide more information that people find useful. During development we’ve already added an additional feature so that when at the lowest level in any force the Find Out More tab gives detailed local information for the neighbourhood policing team.

The site currently focuses on crimes committed, maybe it could extend to allow visitors to report crimes in their area, perhaps it could show how many of the recorded crimes are detected and beyond this show what happens to the culprits?

If you have any suggestions, or any questions, please post a comment or email us.