Refining the Model

Next, we wanted to refine the model to make it as accurate as possible.

My team members worked on finding bus transfer rates, and I worked on finding bus occupancy rates (how many people are on a bus at any time).

To do this I analyzed Go-Card data for the month of April in 2013 that had been provided to us by TransLink. This contained all the transactions of passengers riding the city buses during that month who had used Go-Cards to pay for their fare. There are 1,371 bus routes in Brisbane, each with multiple services that run, so the processing of this information took a long time. I first divided up the file into separate files for each bus route so that it would be easier to analyze them individually. Then I used MatLab to analyze the data for each route and generate a graph of the average number of passengers on a service for each minute of the day. An example from bus route 60 is below.

60graph