Map time (2017)

[nitory]

A simple map of which constituencies which changed parties from 2015.

[Sibboleth]

[Deleted]

Jun 13, 2017 13:53:19 GMT Sibboleth said:

Some islands group in the Caribbean?

[John Chanin]

I've figured it out as Yorkshire. But Sibboleth it would help if you label your maps.

[Pete Whitehead]

Jun 13, 2017 14:14:59 GMT John Chanin said:

I've figured it out as Yorkshire. But Sibboleth it would help if you label your maps.

Nonsense the outlines are instantly recognisable to everyone here

[John Chanin]

It took me 3 minutes or so to figure it out. I'm a numbers person not a visual person.

[clyde1998]

Using the ward level results on the Electoral Calculus website, here's a ward map of Berkshire (darker colours represent a larger majority - upto 50%):

[Pete Whitehead]

Jun 13, 2017 16:19:01 GMT clyde1998 said:

Using the ward level results on the Electoral Calculus website, here's a ward map of Berkshire (darker colours represent a larger majority - upto 50%):

On past experience I'd take the ward results from EC with a huge dose of salt, but i'd be interested to see what they have - do you have a link to their figures? I'd like to see how they compare with what I've done so far in Hertfordshire

[nitory]

Jun 13, 2017 17:01:01 GMT Pete Whitehead said:

On past experience I'd take the ward results from EC with a huge dose of salt, but i'd be interested to see what they have - do you have a link to their figures? I'd like to see how they compare with what I've done so far in Hertfordshire

You can look up their constituency results by ward here.

Belfast West breakdown is always worth a giggle... DUP winning the Falls ward.

[clyde1998]

Jun 13, 2017 17:01:01 GMT Pete Whitehead said:

Jun 13, 2017 16:19:01 GMT clyde1998 said:

Using the ward level results on the Electoral Calculus website, here's a ward map of Berkshire (darker colours represent a larger majority - upto 50%):

On past experience I'd take the ward results from EC with a huge dose of salt, but i'd be interested to see what they have - do you have a link to their figures? I'd like to see how they compare with what I've done so far in Hertfordshire

Yeah - some of them do look off, tbh. That would probably be down to them not directly taking into account local issues - just local elections results. The local election results for West Berkshire, as an example, are quite hard to port into general election results - as, for most wards, there are only Lib Dem and Conservative candidates. My ward is one of the few to have a "split result" - ie one Conservative, one Lib Dem. That seems to be down to a popular Lib Dem candidate - rather than any above average support for the Lib Dems - although I'm sure it's very hard to see what wards have a very popular candidate in, if you don't live there. Although, this is the only ward dataset that I can find - without making my own, of course (which is something I'm looking to do).

The ward results are found in each constituency's page - you'll have to search here. Their ward estimates are on the table at the bottom of the constituency's page. Once they start adding opinion poll data in to their predictions, the estimated results are the final table under workings tab (underneath the ward-by-ward heading).

[ccoleman]

According to Electoral Calculus, Labour were just around 300 votes behind the Tories in the Alwoodley word in Leeds. To anyone who lives in or around Leeds they'll know that is quite remarkable if even remotely true.

[clyde1998]

I sent the guy that runs that site an email about whether he'd be willing to send me his estimates for the ward level results from the 2015 election. The response I got back was:

Thank you for your interest in Electoral Calculus and the data.

The ward level data is a proprietary dataset, which has been carefully put together from a number of sources, including local election results and data regression against census demographics, combined with general election results. It includes party support, as well as EU referendum vote, and political attitudes.

The dataset is available for licensing on commercial terms. Please let me know if you are interested and I would be delighted to help your project with our data.

I'm tempted, as it could be quite interesting to see where some of these results come from. I might see if he'd be able to provide the data on 'research' terms, as "commercial terms" is probably well above my financial range and probably isn't really the terms of use I'm looking to use their data for.

If others are interested, I'd be willing to send a response asking about the terms of cost of the commercial license and to see if he'd do it on different terms if it's completely unrealistic for a private individual could afford and would want to use the data for.

[clyde1998]

I've previously saved the estimated ward results for my local council (West Berkshire Council; 30 wards), which is good as it allows me to see the difference between this election and the last in some wards.

It's clear that it's not uniform swing. The median change is votes in the wards in question are Con -0.4%; Lab +5.9%; Lib +5.6%, with a median swing to Labour of 3.3% from the Tories.

There's one ward that sticks out like a sore thumb. For no apparent reason, Calcot (which borders Reading, should be in the Reading Borough Council area and is in the Reading West constituency) had a whopping 26.2% swing to Labour, when the next highest was 6.7%. I'd love to know if that really happened or not.

[Pete Whitehead]

From what I've looked at the ward results look a lot more plausible than previous efforts though I've only really looked at Hertfordshire and London seats

[Lord Twaddleford]

Interesting tool there. I have to question some of the conclusions it comes up with, but looking at the breakdown for my own constituency most of its conclusions seem plausible, though a few of them seem questionable to me.

Hmm. Combine this with the Boundary Assistant tool, and let's see what kind of notionals one get using these conclusions...

[lancastrian]

Jun 13, 2017 18:52:57 GMT Pete Whitehead said:

From what I've looked at the ward results look a lot more plausible than previous efforts though I've only really looked at Hertfordshire and London seats

Looking at Lancashire, they generally look okay-ish, but Lancaster & Fleetwood, especially the university, looks off, and some of the electorates in Burnley and Hyndburn look significantly off. Issues remain with wards not properly contested in local elections.

They're obviously going to struggle with some of the big swings seen in this election - see Canterbury.

[Clarko]

They're way off with some of the GM seats - particularly impressed they've managed to allocate 19 votes to the Greens in Heywood and Middleton where they didn't have a candidate.

[Harry Hayfield]

Whilst the arguments are sound (which sounds like something straight out of a Gilbert and Sullivan comic opera) some of the results I am not convinced by. For instance here in Ceredigion, this is the change in my ward between 2015 and 2017

Llansantffraed 2015: Lib Dem 35%, Lab 21%, UKIP 20%, Con 13%, Green 10%, Plaid 0%, Others 2% (note of concern: There wasn't an Other in 2015)
Llansantffraed 2017: Lab 30% (+9%), Con 25% (+12%), Lib Dem 23% (-13%), Plaid 16% (+16%), UKIP 3% (-17%), Green 2% (-8%), Others 1% (-1%)
Labour GAIN from Liberal Democrat on a swing of 11% from Lib Dem to Lab
Swing in constituency as a whole: 8.7% from Lib Dem to Lab

[clyde1998]

Jun 13, 2017 18:21:33 GMT clyde1998 said:

I sent the guy that runs that site an email about whether he'd be willing to send me his estimates for the ward level results from the 2015 election. The response I got back was:

Thank you for your interest in Electoral Calculus and the data.

The ward level data is a proprietary dataset, which has been carefully put together from a number of sources, including local election results and data regression against census demographics, combined with general election results. It includes party support, as well as EU referendum vote, and political attitudes.

The dataset is available for licensing on commercial terms. Please let me know if you are interested and I would be delighted to help your project with our data.

I'm tempted, as it could be quite interesting to see where some of these results come from. I might see if he'd be able to provide the data on 'research' terms, as "commercial terms" is probably well above my financial range and probably isn't really the terms of use I'm looking to use their data for.

If others are interested, I'd be willing to send a response asking about the terms of cost of the commercial license and to see if he'd do it on different terms if it's completely unrealistic for a private individual could afford and would want to use the data for.

I've worked out the cost of getting the entire country's data (wards + output areas) is around £65,250, based on the information I've been sent. That's a bit more that I'd be happy to send on this sort of thing.

Even for my council area, it's £750 (£25 per ward) - which is good to know if I ever run for an election and I have a budget to work with. 

[Deleted]

Would be good if people posted maps rather than text!