The data provided by Fairtiq allow for a deep data analysis. In order to reduce the commuting time of people every day we focus on the most annoying part of the journey: the transfer waitings. Our goal is to analyse the waiting time in transfer stations and propose solutions in order to reduce it.

What it does

It analyses every transfer journey made during a standard week in the canton of Fribourg/Freiburg (CH). It provides basic statistics such as number of total transfers per hour or average waiting time during the day. Statistics for each transfer station are also computed and afterwards plotted in a map. Each station in the map provides the critic period (meaning the time where more people wait longer times) and also the most frequent transfer journeys and the average waiting time for each of them. This data allows for a better understanding of the current situation. From the diagnostic, the measures that can be implemented in the transport network in order to improve the efficiency of the system are tackled.

How we built it

We analysed the transfer journeys from the available data from Fairtiq. As starting point, we used Matlab for analysing the journeys, identifying the critical transfers in the network in terms of the time spent by the users while transferring and the amount of transfers occurring daily in each station. The stations have been ranked by applying machine learning techniques for clustering and classification. For the sake of visualisation, an interactive map has been developed using javaScript and Google's map api, plotting some of the most significative critical stations. Finally, optimisation of the most crowded and longest transfers for those stations have been further studied and possible solutions for improving the user experience have been discussed.

Challenges we ran into

Data analysis has been very challenging due to the huge amount of data available that was not directly applicable for our goal. So, understanding the data set and, above all, the extraction of useful features and statistics have required a lot of effort. Once the data has been cleaned, the handling of data by applying different approaches in order to get our rankings has been challenging too. Finally, the request of coordinates of stations via the Google Maps API has required to work with new data fetching platforms.

Accomplishments that we're proud of

We are proud of having built a model providing information that we strongly believe could be useful for the improvement of commuters' time and efficiency of the transportation network.

What we learned

A lot!! Mainly dealing with huge amount of data, selecting the useful and visualising it properly!

What's next for Travel more, Wait less

The next steps are to try the model with data from bigger cities and regions. We believe that transfer times could be much more important in cities with higher density of people and with more intra city lines.

Share this project: