The inspiration for doing this piece of work was looking at the chloropleth map of Welsh local government on wikipedia. Colouring in constituencies works reasonably well at a Westminster / Welsh Assembly level to show where parties are strong where. But once you get down to local government ward level it can be very misleading, for example Blaen Hafren, a single member ward in Powys, is 266 km2 considerably bigger than the whole of Cardiff (140 km2).
I looked at alternatives and decided that adding dots for each councillor (dot placed within ward) would give a better idea of how strong parties are across the country. Since I would need to get the party of each councillor I also added names, email addresses and links to the council websites.
This is the output of that analysis. You can get the councillors details by clicking on the dot. You can also select the parties to display and open the map in fullscreen mode.
It is instantly striking that the parties have very different spreads of support. Labour have no councillors whatsoever in Mid Wales and are generally weak in the west of Wales. Plaid Cymru are strong in the West but also have considerable support in some of the south Wales Valleys. Conservatives do poorly in West Wales outside of Pembrokeshire. Liberal Democrats are a strong force in Cardiff and Powys and have a reasonable presence in Swansea and Ceredigion but are weak elsewhere. Independents are a massive force across the whole of Wales.
As well as looking at the geographic location of councillors we can also compare councillors by some of the characteristics of their ward. For this piece I have chosen three characteristics.
1. The percentage of Welsh speakers in a ward (2011 census)
2. The median house price in a ward (September 2016 to September 2017)
3. The percentage of Welsh identifiers in a ward (2011 census)
Each of the graphs below plot these variables against each other. (You can select/deselect parties on the legend and can hover over the points to get councillor details)
If you try comparing just Labour and Conservative you see (unsurprisingly) that the Conservative wards tend to have much higher Median house prices whilst Labour tend to be lower. If you add in Plaid you notice that wards with more Welsh speakers tend to be Plaid.
It is noticeable here that there is less obvious differenece between Plaid and Labour wards but that Conservative wards tend to be less Welsh identifying.
This final plot clearly shows that Plaid wards tend to either have a high percentage of welsh speakers or a high percentage of Welsh identifiers.
At one level this analysis is nothing new, Plaid are strong where people speak Welsh or identify as Welsh. Labour are strong in Welsh identifying areas and poorer areas. Conservatives are strong where people are wealthier. No great suprises.
However one of the interesting elements of this analysis is that it also lets us identify councillors who have been elected in wards which are not typical for their party. Sion Jones (Bethel, Gwynedd) is the only Labour councillor in a ward which is over 70% welsh speaking. In fact there are only a handful of Labour councillors in wards where more than 50% are Welsh speaking. What are these councillors doing that others are failing to achieve?
Likewise Plaid have won Menai-Bangor (Gwynedd) and Aberystwyth North (Ceredigion) despite low Welsh speaking and Welsh identifying percentages, what can these councillors (Catrin Wager, Mair Rowlands and Mark Strong) and the teams that got them elected teach their party?
And for the Conservatives can they learn anything from successes in more deprived areas such as Rhyl East (Tony Thomas) and Rhyl South East (Brian Jones) and Welsh identifying areas such as Mawr (Brigitte Rowlands) and Ton-Teg (Lewis Hooper).
It might be that these results are down to very local circumstances, well known candidates or poor campaigns by other parties. But it is also possible that these councillors have got important messages about how their parties can appeal to new and different voters.