when parliament runs an inquiry about you…
in 2018 the victorian parliament ran an inquiry into whether vicroads had been effective in managing victoria’s regional road network…
As parliaments do, this was accompanied by a submissions process, where any interested stakeholder is welcome to write a letter to the Parliamentary committee for their consideration. Altometer working with VicRoads was able to demonstrate the power of data science and machine assisted data analysis techniques as a way to quickly make sense of what had been said.
In all around 350 submissions were received by the Parliament and posted to the committee website. VicRoads, the agency responsible for the operation of the state’s road network, wanted to understand who had written to the parliament and what was their point of view.
VicRoads and Altometer worked together to analyse the 350 lengthy submissions. They were mostly very well considered by the Victorian community, spreading over multiple pages and expressing a nuanced views on VicRoads performance.
Altometer used this process as a test for a machine learning based technique for qualitative data analysis. The analysis would be run in the traditional way, with a team of people reading, coding, categorising and reporting on all the topics raised among all 350 responses. The comparator would be using a machine assisted process that allowed a machine to sort and categorise and human analyst validate their work.
The results were astonishing.
For the same 350 responses, both techniques were able to identify 5 major topics, each with a range of issues within them - some also analysed for sentiment.
However the work, and hence the time, to complete one analysis over another was reduced by 75%. This is the equivalent of an 80 hour process (two people working full-time for a week) reduced to a 20 hour process (or one person for half a week). All for the same result.
If you want to know more about how data science and machine learning can benefit your organisation: