2022 Election Dashboard
This dashboard was built using Plotly's Dash written in Python. The data were sourced from the Australian Election Commission for elections variables; the Australian Bureau of Statistics for census variables, boundary shapefiles, and correspondence tables; and the Parliamentary Library for opinion polling history with the exception of the 46th Parliament which was scraped from Wikipedia using Beautiful Soup.
Election result variables included:
- First preference (1P) votes by polling place
- Two candidate preferred (2CP) results by polling place
- Two party preferred (2PP) results by state
- Distribution of preferences, by polling place
- Nominations by gender
- Number of nominations
- Pre-poll and postal vote counts per day by polling place
- Total vote count by SA1, polling place, and vote type
Elections results included elections between 2004 and 2019 at the polling place, electorate, state, and national levels. For elections between 2013 and 2019, results were also converted to the SA1 and postcode level and converted into the 2022 electoral boundaries for comparison across years. This was achieved with a combination of the AEC's total vote count by SA1 spreadsheet and the ABS's correspondence tables.
In addition, 2022 pre-poll and postal vote counts were sourced from the AEC website which are updated daily during an election campaign.
Census variables included:
- Over 4000 variables, including demographics, employment, family, internal migration
Census variables included the 2011 and 2016 censuses at the SA1 level and lowest possible answer-level for each. 2011 census variables were converted into 2016 boundaries for comparison across years. Like the election conversions, the ABS's correspondence tables were used.
Election Dashboard:
Overview
The overview section of the dashboard provides a high-level overview of key electoral information over the six elections (2004 - 2019). It provides a quick view of how the vote of each party has progressed over time.
Features include:
- Total seat counts, two party preferred (2PP) vote, and first preference vote (1P) by election and party, including the change in these variables between elections
- Ability to filter by state and isolate single parties
Voting
The voting section provides an update on voting prior to election day throughout the campaign. This provides an understanding of how many voters have already voted and the history of votes by voting type so that campaigns know where and when resources are needed.
Features include:
[First graph]
- Compare across multiple elections to see how the 2022 election compares to previous elections on the same campaign day.
- View counts of pre-poll and postal votes daily or as a cumulative figure (over actual turnout or enrolment) to track how many voters have already voted prior to election day.
- Filter by state, electorate, or polling place.
- Update 2022 vote counts automatically and in real time.
[Second graphs]
- Compare the proportion and number of 1P (including informal votes or without) and 2CP (classic and non-classic) votes by party and vote type (incl. ordinary, pre-poll, postal, absent, and provisional) to understand where a party traditional performs better (e.g., the Coalition tends to perform better in postal votes than ordinary votes).
- Compare the proportion and number of 1P and 2CP votes by year and vote type to understand where the majority of a parties votes are likely to come from (e.g., on election day or before it).
- Filter by state, electorate, or polling place to have a more granular understanding of where resources are needed based on previous elections.
Preference flow
The preference flow section shows the distribution of preferences. This can aid in the preparation of 'how to vote' cards by giving an understanding of how voters preferenced at previous elections.
Features include:
- A visual representation of preference flow using a Sankey diagram and also in a stacked bar chart for the proportion of votes distributed from and to parties, and a bar chart for the number of preferences distributed from and to parties. This helps to quickly see where votes were distributed to in both a classic (2PP) and non-classic (2CP) contest.
- Filter by year, vote type, state, electorate, and polling place.
Pendulum
The pendulum section quickly shows where the marginal/key seats sit on the post-election pendulum.
Features include:
- Compare by party of year
- Filter by state
Candidates
The candidate section provides information on the gender divide of candidates and number of nominations. Gender is a typical question that is raised during a campaign.
Features include:
[First graph]
- Compare the proportion of male vs female candidates by party or year
- Filter by state
[Second graph]
- View the distribution of the number of nominations by year
- Filter by state or area type (e.g., inner metropolitan vs rural)
Opinion Polling
The opinion polling section provides historic opinion polling results. Media often comments in regards to polls. Having the historic results can therefore be useful.
Features include:
- 1P and 2PP voting intention between 2001 and 2022, including key events (such as a leadership change) and campaign periods.
- Preferred PM and Satisfaction ratings of leaders between 1987 and 2022.
- Ability to isolate parties and zoom in on certain years.
- Filter by pollster and interview type (e.g., CATI, online, etc).
- Update with new polling results.
Electorates:
The electorate section provides an overview of all census variables and a few key headline figures for Australia, each state, and all electorates. This is useful in gaining a quick and deep understanding of the electorate as a whole.
Features include:
- Converts 2013-2019 headline election results (informal vote, Coalition first preference vote, proportion of preference flow to the Coalition) into the selected boundary level and also provides a visual aid of the 2CP result for these elections so that a trend can be seen.
- Provides over 4000 census variables including two visual aids of a population profile and index of relative socio-economic disadvantage.
- The remaining variables are in word and table format under the headings of Demographics; Employment, Income and Education; Dwelling and Household Characteristics; Family Characteristics; and Internal Migration.
- For ease, all variables are ranked from the highest proportion to the lowest under each category. They also appear as either red for being below the Australian average or green, for being above.
- Data can be converted into 2022 or 2019 electoral boundaries for across-year comparisons.
- Filter by state or electorate.
Data Analytics:
Exploratory
The exploratory section provides a quick and easy way to find areas with high proportions of the desired variable(s) by ranking them from highest to lowest.
For example, a first home owners grant scheme announcement may require a campaign to find where there are high levels of renters in a certain age bracket and on a certain income. This could be achieved by adding the three census variables of Tenure Type, Total Family/Personal Income, and Age Groups. The postcode results of this could then be used on social media.
Features included:
- Add as many election or census variables as desired.
- Filter variables by a minimum proportion (e.g., only include electorates with more than 5% of the electorate with people who were born in Scotland)
- Export the table to Microsoft Excel.
- Map the data in MapBox.
- Rank data by 2022 electoral boundaries, 2019 polling places, 2016 postcodes, or 2016 SA1s.
- Filter by state or electorate.
- Option to remove "not applicable", "not stated", "partial income stated", "inadequately described" and/or "overseas visitor" from the denominator.
Regression
The regression section provides the ability to quickly test the relationship between two variables or view the distribution of one variable.
Features included:
- Add as many charts as desired.
- Choose from over 4000 census variables or 2013-2019 election results.
- Colour data points by party, or 2019, 2016 or 2013 2CP Outcome.
- Hover over data points to view which electorate the data point belongs to.
- Interpretation of relationship between two variables provided in an easy to understand way for those with no mathematical background.
- Ability to remove outliers.
- Plot data points as 2022 electoral boundaries, 2019 polling places, 2016 postcodes, or 2016 SA1s.
- Filter by state or electorate.
- Option to remove "not applicable", "not stated", "partial income stated", "inadequately described" and/or "overseas visitor" from the denominator.
