How Theresa lost her majority

[jluk234]

I have been working on a 'little' project, inspired by and done in a similar way to this article, in which I have, for each region, put each seat into a category based on population and how much of each seat is built up and I have calculated the voteshare for the category for every election since '87.

I will go into more detail about this later tonight when I put the first set of results up.

[jluk234]

After coming across this, I have been doing a similar thing for last year's general election. I asked myself a question right at the outset.

Could the Conservatives losing their majority have been foreseen looking at the results for every seat in every election since '87?

With this, I went forward and started categorising the seats in the East of England. I examined each constituency and its local area. As a result, I decided upon four categories.  These are:

• City - Constituencies that cover a part of a city or have a large population and most urban areas that have more than one seat
• Large Town - Constituencies centred on an urban area that has a large population but not large enough to be in the city category
• Town - Constituencies that have several towns that have a reasonable population and numerous other towns
  
• Rural - Constituencies that have their population thinly spread across the towns and have large rural farmland areas
  

[jluk234]

The first bit completed is the city category for the East of England.

These maps show which categories each seat is in for the election between 1987 and 2017.




And these are the results for all those seats in red
 

These are the detailed seat by seat result for each election

[Pete Whitehead]

I like what you're trying to do and don't want to appear negative, but how have you arrived at these classifications? There seems to be a lot of inconsistency about how different constituencies have been allocated. For one thing St Albans isn't a city in anything other than the cathedral sense by which logic you should have SE Cambridgeshire in the same category (due to Ely). In fact but for Ely, SE Cambs is almost entirely a rural seat as is South Cambs - certainly they have less of an urban population that several of the seats in Norfolk and Suffolk that you've classified as rural.
You probably really need to have a greater number of categories anyway (Johnson, Pattie and Allsopp had about 30, though that was excessive and there was a good deal of overlap) but whatever categories you have you've got to be clear about defining what they mean and having clear criteria as to why constituency x should be included in one group rather than another

[Deleted]

It might be hard to quantify, but I wonder if an 'edge of conurbation' category might be appropriate for seats that don't fit easily into 'city', 'town' or 'small town'. It might also be more meaningful than the above for many of the kind of seats that went Labour in 2005 and Conservative in 2017.