By now you’re probably aware of the UK Government petition about Donald Trump. It’s currently be signed by over 500,000 people, making it the most successful petition in the UK…
But who actually hates him most? The site conveniently provides a map, so you can see where people dislike him… but constituencies aren’t all equal, and you always run the risk of doing a heatmap of population densities:
So I thought I’d explore further. Firstly, which are the top constituencies by proportion of the population who hate Donald?
|Bethnal Green and Bow||Rushanara Ali MP||1874||1.50%|
|Bristol West||Thangam Debbonaire MP||1779||1.43%|
|Brighton Pavilion||Caroline Lucas MP||1405||1.36%|
|Hackney South and Shoreditch||Meg Hillier MP||1496||1.27%|
|Islington North||Jeremy Corbyn MP||1284||1.24%|
|Cities of London and Westminster||Rt Hon Mark Field MP||1364||1.24%|
|Glasgow North||Patrick Grady MP||876||1.22%|
|Hackney North and Stoke Newington||Ms Diane Abbott MP||1556||1.22%|
|Islington South and Finsbury||Emily Thornberry MP||1237||1.20%|
|Edinburgh North and Leith||Deidre Brock MP||1279||1.20%|
OK, so we see some of the usual faces there. Jeremy Corbyn, Diane Abbot, Caroline Lucas. Maybe it’s just a political thing. Let’s look at the bottom constituencies (again by the proportion who are keen on the bewigged one):
|Blaenau Gwent||Nick Smith MP||92||0.13%|
|Barnsley East||Michael Dugher MP||119||0.13%|
|Makerfield||Yvonne Fovargue MP||126||0.13%|
|Boston and Skegness||Matt Warman MP||129||0.13%|
|Wentworth and Dearne||Rt Hon John Healey MP||119||0.12%|
|Walsall North||Mr David Winnick MP||117||0.12%|
|Doncaster North||Rt Hon Edward Miliband MP||123||0.12%|
|Easington||Grahame Morris MP||102||0.12%|
|Kingston upon Hull East||Karl Turner MP||112||0.12%|
|Cynon Valley||Rt Hon Ann Clwyd MP||78||0.11%|
Well I wouldn’t (necessarily) have expected to see Ed Miliband and Ann Clwyd’s constituencies there.
So let’s go slightly more data sciency. What is the correlation between Labour vote and proportion hating Donald?
Rather than just looking at that, let’s look at a whole range of stuff:
Woah! What was that? Well that’s the correlations between a whole load of stuff and the percentage who hate Donald. Percentage is the second column (or row), and a blue box means a positive correlation, and a red box a negative one.
So let’s look at some of the more interesting ones, and sort them from highest to lowest. Remember that these are at a Constituency level.
- PopulationDensity 0.64
- Green 0.55
- FulltimeStudent 0.53
- Muslim 0.37
- LD 0.19
- Lab 0.07
- Con -0.11
- Ethnicity White -0.48
- Christian -0.60
- UKIP -0.61
- Retired -0.67
So it would seem that the strongest correlation isn’t with Muslim populations, it’s actually strongest with built up areas, then with 2015 Green voters, then with full time students, and only then with Muslim areas.
And it’s probably not surprising that those lease likely to hate Donald are constituencies with lots of retired people, a strong UKIP presence (some immigration is OK then, it would seem), a large number of people identifying as Christian, and white people.
Taking it one small step further, and running our old friend, Linear regression we see a model like this (Labour and Con voters removed to avoid tons of correlation, LD voters removed out of sympathy):
|Estimate||Std. Error||t value||Pr(>|t|)|
A slightly different story – but looking at the key stand outs: if you want to find a constituency that really hates Donald, first look for Green voters, then a densely populated area, with quite a sizeable Muslim and white community, but keep away from those UKIP voters, especially the retired ones.
*Data for the petition harvested at about noon, 11 December 2015
** Other data from ONS based on the 2011 Census
*** Featured image Donald Trump by Gage Skidmore
**** Northern Ireland doesn’t provide breakdown of petition numbers