Improved traffic management and targeted public transport services are the focus of new algorithms being developed to draw meaning from the mountains of data now available about where we travel, how, and when.
In Melbourne, ongoing changes to train, tram and bus public transport services in the CBD provide an opportunity to test what pedestrian data can reveal about shifts in people’s travel patterns as a result of the service changes.
Professor Chris Leckie at the University of Melbourne’s School of Computing and Information Systems has been working on the research with Public Transport Victoria (PTV).
He uses data from the City of Melbourne’s pedestrian monitors in the Melbourne central business district, which is publicly available from the Victorian Government’s open data website, www.data.vic.gov.au.
We use contrast data mining — comparing last year’s data with this year’s data — trying to detect any significant changes in pedestrian movement, Professor Leckie says. The pedestrian monitors track movements by the hour.
The algorithms being developed for analysis need to be smart enough to detect significant changes, and to eliminate ‘noise’, such as the weather, that might otherwise hide the overall trends when comparing one year’s data to another.
We also generate hypotheses about what is causing the changes, and pass these on to PTV for further analysis using other data sources. He plans to add VicRoads traffic data to the analysis, to identify any concurrent changes in traffic patterns within the city.
In the longer term, Professor Leckie says this kind of analysis could also be used to identify how planned events will affect traffic flow. Predictive modelling could then be used to optimise alternative travel paths, including traffic signalling, as well as identifying alternative transport options.