This post accompanies the lightning talk that I gave at the Leeds Data Mill event – ‘Sport In Numbers – What Makes a Sporting City?’ on the 7th October 2014, although this post does also stand up in its own right… See the slidedeck from the event.
Trafford has a very strong sporting pedigree – Manchester United Football Club, Lancashire County Cricket Club, Altrincham Football Club, and Manchester Phoenix all play home games here. We also have Sale Sharks Rugby Club and Manchester City Football Club who have their training grounds in Trafford, and I reckon I could hit Salford Reds’ Stadium with a stone from the boundary between Trafford and Salford. I’ve also recently learned that Samuel Ryder grew up in Sale. Samuel Ryder donated the Ryder Cup for golf’s
only most interesting competition.
Trafford was also a host Borough for the London 2012 Olympics, is an Ashes venue, and is home to the Greater Manchester Marathon – the flattest marathon course in the UK (useful for personal bests, apparently).
Because of all these organisations and events, Trafford Council decided to join forces with Trafford Partnership to ‘harness the many benefits that sport can bring to local communities’.
To support the Sport and Physical Activity Partnership, to help target resources, and identify needs and opportunities , we decided to use multiple datasets to profile areas in Trafford, creating something that we call ‘The Indices of Sporting Need’ (we actually call it ‘that sport data’, but…), and we definitely need a better name for it.
The index was roughly modelled on the Indices of Multiple Deprivation (though with less statistical rigour applied). As such our ISN was built according to these principles:
- Indicators must be available at Middle Super Output Area Level
- Indicators must measure something which could be affected by increase or reduction in sporting provision
- Where possible, data should be open
- Data should be clearly referenced, and updated regularly
So we started to go through potential indicators, and came up with this list:
- Indices of multiple deprivation score
- Childhood obesity (Reception)
- Childhood obesity (year 6)
- Adult obesity*
- Healthy eating*
- People in not good health
- Male life expectancy
- Female life expectancy
- Participation in sport*
- Male deaths from coronary heart disease
- Female deaths from coronary heart disease
- Binge drinking*
- Substance misuse*
- People in drug and alcohol treatment**
- Anti-social behaviour
- Probation cases**
- Young offenders**
* Denotes the data is a modelled dataset
** Denotes the indicator is locally calculated, not nationally available
These indicators were fed into a spreadsheet.
The minimum and maximum values for each indicator were calculated, and became 0 and 100. All other values were then converted to a centile score. This gave a coefficient for each indicator that we were able to colour accordingly. We also totalled those scores, to give an overall coefficient.
This overall coefficient was used as the defining value for the choropleth map.
Once the coloured map was ready, and we had the patchwork spreadsheet, we invited the sports foundations in Trafford in to see outcome, and get a feel for whether they felt the model was a good one, and whether the story that the data told felt right. Thankfully, they liked it…
We then added a layer of our known sporting assets to the choropleth, to give an idea of the sorts of sporting provision that already exists.
What was interesting was that in Partington, a town in the West, red MSOA, there is a cluster of points, showing a leisure centre, a sports village, football pitches etc. And yet there is still, according to the model, need. We decided to take this to the Locality Partnership responsible for the West of Trafford, and it would appear that the local population do not use the facilities. The challenge, then, for the Locality partnership is to work out why this is the case, and how we can better connect residents with the available facilities – which is what they are now working on. An example of using open data to inform partners, and change communities.
Feel free to have a look at the interactive version of the map.
We think that there are some real opportunities to develop this model, make it a bit more robust, and scale it, across Greater Manchester, or beyond! Following the presentation at Leeds Data Mill’s Sport in Numbers event, we have had discussions with Leeds Data Mill, and with Devon County Council, about scaling the model. As a group, this would be a good test – with Trafford (sort of suburban), Leeds (Urban), and Devon (Rural), different issues will present, such as overcrowding vs distance from a facility.
We have also had early discussions with Greater Manchester and West Yorkshire
Community County Sports Partnerships, about potentially using the model to aid priority-setting, and understanding the communities they work in.
We also need to review the indicators that were used to inform the model. I think that there are too many modelled indicators in the list, and I suspect that a proper statto would want to apply weightings to the individual coefficients.
Finally, there are probably things we could do to link to data held by Sport England, and the National Sports Governing Bodies that could really give an amazing view of sporting opportunities, and where they can make a difference.
I think that the important thing to note with this whole concept is that it isn’t perfect, and we know that, but we have started it, and made something that we can build on (hopefully in partnership with others). If you have read this, and want to know more, or get involved, the project will be added to Pipeline, the LocalGovDigital collaboration tool. Or get in touch with me directly, and we can keep building it up. Or, if you know of something that’s already happening in this space, again let us know, and we can might be able to hook it up.