The Spatial Smoking Gun (part 2)
This is the second of two posts about some spatial analysis aimed at reducing adolescent smoking rates. The first talked about the analysis itself while this one discusses how the results were shared with colleagues and end users.
In part 1 of this post, the analysis issues were covered and we ended up with a spatial database of tobacco retailers and their proximity to secondary schools, using geocoding and spatial joins. Our next issue was how best to share these results. There were two aspects to this: sharing preliminary results with my colleagues during analysis and then sharing the final results with our end users (the DHBs). With my colleague, Louise, in Dunedin, it wasn’t going to be easy to sit down with her to go over the maps. For a while I would email her screen shots of what I was working on, but that wasn’t ideal either when she wanted to look at specific schools – lots of tooing and froing. The primary end users of our results were the district health boards and one of the things they wanted was maps of the outputs. And therein lay our first challenge. Of the 20 DHBs, let’s start by looking at the smallest one, Capital and Coast, to illustrate. Here’s a map showing the area that this DHB covers:
Within this district are 24 secondary schools and 324 tobacco retailers. How best to show these on a map? There’s a big scale issue at work here which became clear when we starting thinking about hard copy maps, and one that gets more problematic as the DHBs get larger. How much useful information would one map of the whole district convey? We could show the locations of all the schools, but at the scale of the whole district, it would be impossible to see the walking zones around each school. Adding the retailers would just make things messier. So maybe we could make a separate map for each school in each district? Let’s see, that would be 20 separate district maps and as many as 478 individual maps for each school. In urban areas we could try and get away with several schools per map, but at a cost of using a scale that might be too coarse. We could set up a map template to make life easier but we’d still be dealing with a lot of maps and coordination to keep things consistent. The thought filled me with dread, to be honest.
So while we could have done paper maps, a much better route for all of us was to use web maps. With webmaps, Louise could see all of the latest results from the comfort of her own computer or tablet from wherever she was (yes, even her phone), so long as she had a good internet connection. For the DHBs, webmaps enabled them to easily get as much detail as they needed rather than having to rely on pre-made paper maps. Using the Snipping Tool in Windows they could put together their own figures for reports (as Hawkes Bay have been doing).
So for Louise, I used the webmap below to share results during the analysis. I would do my analysis on my desktop machine, then upload those data to the GIS server, create the webmap and send her the link. She would ooh and ahh (if I was lucky) and give me feedback on what she wanted to do next (click on the image to go to the map).
For the clients, we needed to set up a webmap for each of the DHBs, so a template was created to keep all of the maps consistent in style and content. Once the data were on the server, we developed a web application, customised to each DHB and then sent each the link to their particular district (click on the image to go to the map).
We set up pop up windows so when someone clicks on a school or a retailer, a table opens showing the attributes for that feature, so it provides a little look into the attribute table.
Being a webmap, what the end users are actually seeing is an image rather than actual data. So while they can see what’s there, pan around, and turn layers on and off, they can’t do any real analysis with the data. For that we would either have to send them copies of the data (which we’ve done for Hawkes Bay) or set up a web feature service to they could work with some actual data using ArcMap on their end.
So far, this has been working out well for the DHBs and they’re actively using these results for their work. Being able to share my preliminary results with Louise was made a lot easier by setting up the webmap so she could see what I was seeing.
The process of setting up a webmap isn’t all that difficult and I’ll cover it in another post. If nothing else, this post may spur some ideas for how you might be able to share some of your mapped results with others. If so, give us a ring.