open doors: 130,000 public responses
in september 2018, the QLD Government engaged queenslanders on the best ways to improve renting in queensland
The number of responses received was tremendous: over 130,000 responses were collected in a range of formats. With a policy decision needed quickly, machine assisted analysis techniques were used to cut analysis time to under 3 weeks.
Altometer teamed up with ARTD, government evaluation specialists, to get to work on the data. Using machine assisted data analysis processing time was cut by 75%, making way for a report for cabinet’s consideration.
Making the process more complex was the fact that the data came in several formats:
Postcards (handwritten)
Online surveys
Written submissions (including email)
Forum posts
Social media posts
Snap polls
In all there were over 42,000 qualitative responses to review. All of these were focused on the simple question: What are your ideas for improving the experience of renting? The ranged in length from a few sentences to pages of detailed responses.
The largest swathes of qualitative data were analysed using machine assisted data analysis techniques, where machines read, sort and categorise comments for human analysts to then review and validate.
As one would imagine, the range of responses to this were vast. Although almost 80% of responses were from tenants, a further 18% came from landlords, both with significantly different points of view. Topics included:
Permission to have pets at a rental property
Making minor alterations to a property
Housing standards
Repairs and maintenance
Lease arrangements
Notice to vacate
Property managers
Dispute resolution
Affordability
The team were able to identify 10 main topics, indicate how many times each topic was raised and by whom. This provided the Department and the QLD Government with the information it needed to continue a dialogue with renters and landlords toward what they can do to improve renting in the coming years.
For more information on the Open Doors program, take a look at their engagement site.
For more information about using machine learning support to speed up your data analysis process: