Thank you to everyone who participated in Melbourne’s first Datathon for Social Good!

After attending the launch, Dr Sonja Hood, CEO of Community Hubs Australia (CHA), said that the “success of the weekend floored me.”

During the subsequent working time, the COVID-19 crisis enveloped Australia. Amazingly, despite the stress and challenges presented by this, nine teams pulled through to contribute a submission for CHA.

The datathon attracted a diverse range of people, with over 70% of participants identifying as female. Participants said they were proud to learn and practice skills, and gained a “sense of accomplishment”.


Learn more about the motivation behind #dfsg2020:


For more information about the internal and external tracks pursued by participants, read the problem statements.

All teams have released their work under the terms of the MIT license.

The external track projects include PHIDU content, released under a CC BY-NC-SA 3.0 AU license.

Internal Track

Little Big Stories

Built an interactive dashboard profiling each hub by program activities and referrals.


Analysed program activities and service referrals.

Standard Deviants

Looked at typical and outlier hubs and built an interactive dashboard.

Hit N' Run

Assigned personas to hubs based on program activities. Looked at commonalities across all hubs.

Team Sunshine

Created a report explaining key findings regarding program activities. Built a model to predict a hub's future performance based on attendance.

External Track

Data Wizards

Built a predictive model to suggest which LGAs to open new hubs.

Curious Zucchini

Built a predictive model to suggest which schools to open new hubs.


Analysed LGAs with and without existing hubs. Created suburb profiles.

Geo Hubs

Suggested LGAs to open new hubs based on a set of criteria and plotted them on a map.

  • Download the webpage showing a map of recommended LGAs
  • Download the webpage showing a map of LGAs that already contain hubs
  • Folder containing code

Individual submission

Used look-alike models to suggest which schools to open new hubs.