CPC Predictive and Prescriptive Analytics

This project is a cross-collaboration between research and operations to generate a report that can be run to identify causal relationships and predict the requirements of the Charles Perkins Centre (CPC) hub moving into the future. This involved the Extraction, Transfer and Loading (ETL) of several large data sources, visualisations and data analysis.

CPCData Project

Project Owner

  • Nicole Hoare (Executive Sponsor, ICT)
  • Michael Milne (Business Unit Sponsor, CPC)
  • Quentin Parisotto (Business Unit Sponsor, CIS)

Project Credits

  • Jim Cook (ICT, TechLab)
  • Sonya Corcoran (ICT, TechLab)
  • Chloe (Yixuan Zhang) (ICT, TechLab)
  • Aaquib Ladiwala (ICT, TechLab)
  • Lydia Gu (ICT, TechLab)

Start Date

21 Aug 2017


Completed 04 Oct 2017        

Problem Statement

The Project team are expecting the data to help answer the following questions:

  • Where are good ideas going to come from?
  • What is the predicted capacity for opportunity?
  • How do we build new Multi Discipline Research Centres (MDRC) to help them achieve in future?
  • How much will it cost to run the CPC into the future?

Final Brief

  • A documented and reusable process of performing ETL on the identified data sources.
  • A Power BI integration (or other) of the cleaned data to allow for self guided interrogation of the data sources.
  • A documented process of the classification of the targeted variables within the data sources.
  • Descriptions and justification for selection of modelling and statistical techniques.
  • Recommendations for proceeding and improving/integrating this process further. Provide first report before 20th September to allow for Nicole to present the project at Big Data & Analytics Innovation Summit.

Challenges & Learnings

  • C1. Communication breakdown with new casual hires resulted in duplication of workload and undue stress.
  • L1a. A consistent communication framework needs to be implemented to allow for grievances to be aired constructively.
  • L1b. Clear allocation of duties, workload and regular team updates.

  • C2. Project owners unsure of privacy and consent requirements across their individual datasets.

  • L2. Ensure University privacy policy is communicated and understood prior to commencing project.

Languages / Framework

Python, R, HTML, JavaScript, CSS, Spark, Azure Machine Learning Studio, Azure Blob Storage, PowerBI

Links to Resources