Name: uk-elections
Owner: Data for Democracy
Description: Identifying insights and patterns from analysis of UK election data.
Created: 2017-05-13 10:09:12.0
Updated: 2017-05-21 18:21:15.0
Pushed: 2017-06-20 18:54:46.0
Homepage: null
Size: 476
Language: null
GitHub Committers
User | Most Recent Commit | # Commits |
---|
Other Committers
User | Most Recent Commit | # Commits |
---|
Project Description: This is a project set up to identify insights from analysis of UK election data.
Members (Slack): @james-sr @lindaduong @ktloh
Slack channel: #uk-elections
Project Goals:
Short-term Focus: Given the upcoming General Elections on 8th June, this project will have a short term focus of analysing the historical general elections results in the UK with the aim of gaining insights from past data and conducting predictive modelling on the June elections.
Long-term Goal: The aim is to provide a comprehensive suite of results and variables, both local and national, to construct a unique source of elections data for the UK. Currently election results, stats and analysis are stored in various data sources and reports which adds a layer of complexity for the public to parse through and gain useful insights. This project seeks to address this gap.
Getting started:
The UK general elections analysis phase of this project will follow the general outline below. Details will be listed in the Issues/Projects (General Elections 2017) pages.
Stages:
1. Data collection: Collecting historical general elections results and supporting data, starting with 2015 then working our way back in time.
2. Exploratory Data Analysis: Identify variables to help with prediction and modelling of future/upcoming elections.
3. Prediction/Modelling: Explore different modelling options, such as a binary model that predicts the winner, with a degree of probability, using the 2015 model as a base/train model (potentially add the 2010 data to have more data to work with), or predicting a number of votes per party at constituency level.
4. Data Visualisation: Consider ways of displaying results, e.g. build dashboards, interactive visuals on a site, etc.