Workers, adventurers, explorers: uncovering activity patterns in Melbourne

The bulk of transport modelling is based on peak hour travel and focuses on the daily commute. However, this only considers when and where people are, and also ignores the growing travel demand outside peak hours. Understanding the nature of activities people are participating in is also important in order to understand daily patterns and changing behaviour. Using daily activity-travel data provided by 16,000 households and 41,000 individuals in Melbourne and regional Victoria in 2009-10, this paper uses the k-means clustering technique to uncover similar groups of individuals. Both weekday and weekend activities are explored. This is also combined with socio-demographic data to identify different types of behaviours. Although there is still support for a three-category classification on weekdays (work, education, other) and a two-category classification on weekends (home, other), extending the number of clusters to seven and six respectively allows more detail on the larger "other" clusters. In particular, we can look at the different socio-demographic details dominant in each cluster to determine patterns. More differences between regions were expected, however this was not apparent from the clustering. Some differences, in that those who live in inner Melbourne were more likely to be working or socialising in the evening, were seen, but these were rare. The results contribute to increasing our understanding of how people move around Melbourne, by providing more detail with regard to where and when people are travelling and what people are doing over the course of a day. Future work involves replicating the process for Brisbane and Sydney and comparing the outcomes, leading to better understanding of activity and travel behaviour for Australian cities