Skip to content

Latest commit

 

History

History
32 lines (24 loc) · 2.71 KB

Pitch talking points.md

File metadata and controls

32 lines (24 loc) · 2.71 KB
Raw

1. Data science problem

I am a concerned taxpayer and an engaged voter. How can I inform myself about how the government is spending its budget? Is it allocating capital farily according to societal need and to maximise the benefit to people? Or do electoral pressures mean that some places get more than their share - e.g. marginal electorates? How can I map the dozens of government projects, worth billions of dollars, across 93 electorates (in Queensland state government alone), to keep an eye out for pork barreling?

2. Datasets we'll use

Core:

  • Electoral boundaries. Available as shape files on QSpatial website for 2017 boundaries.
  • Electoral results. (To determine safe ALP / safe LNP / marginal for each electorate.) Available on ECQ website by electorate for 2017 and previous elections.
  • Queensland State Infrastrucutre Plan. (For spending.) Available from DSDTI website as pdf document.
  • Census data, such as population growth and demographics (indicators of public needs in health care, child/aged care, public transport, road), private/public business growth (indicator of building/road/port construction requirement), income growth

Enrichment:

  • Tender data. (E.g. to identify contractors/party patterns.)
  • Stated policy objectives of government: e.g. declared priority development areas, Office of Northern Australia plans.
  • Others we identify as we go.

3. What we'll do

  • Integrate the spending data and the census data. We'll build a model to predict the likelihood of getting a budget approval for an electorate based on Census data. Based on the model, we can test our hypothesis of whether the government spent money based on public needs.
  • Ingest the electoral results, categories into safe ALP / safe LNP / marginal. We'll also confirm this hypothesis that the budget was more likely to be allocated to a certain electoral for political influence
  • Visualise the results, see what we find. Spatial exploratory data analysis.
  • Mix in from the enrichment data sources.

4. How we'll tell the story

  • Build a user friendly web tool allowing citizens to explore/visualise government spending and any spatial relationships to electoral status, in the context of other possible explanations for variation.
  • We may look broadly e.g. Queensland and focus our enrichment effort on a narrower area e.g. SEQ or Brisbane.

Feedback received from Professor Shazia Sadiq and Dr Thomas Taimre on pitch via Zoom 16/09/20

  • Data ingestion techniques such as OCR (optical character recognition) can be frustrating. Be careful of time management.
  • Remember that, in the end, the product needs to be something visual.
  • Making a web tool could be too ambitious / time consuming. Don't be afraid to settle for some static visualisations.