Every day, trains waste very precious time at the stations, waiting for passengers to board or leave the train. Sometimes, that time is affected by circumstances beyond the control of the driver and the passengers. Our goal is, therefore, to help identify the causes of delay, so they can be addressed in order to improve customer satisfaction and overall efficiency.

What it does

traffiK works by analysing the data provided by the SWR in order to provide a complete visualisation tool that can be used by operational staff. Its aim is to identify and highlight the key places and times where improvement may be needed. It also provides more advanced data, such as CCTV footage and report logs to allow more in-depth analysis to understand why trains are exceeding their dwell times.

How we built it

The visualisation app utilises an integrated approach: we developed it using a number of open-source tools meaning it is a low-cost, adaptable solution. The app front-end is based on R-Shiny with javascript, leaflet, HTML and CSS. The data processing was carried out using Python for data wrangling and machine learning (not included as the model was overfitting due to lack of data) as well as R for the plots.

Challenges we ran into

Data cleansing was particularly challenging: the format of the data was difficult to manipulate and required pre-processing before we could get to the machine learning part. Designing a tool that would show complex information in an intuitive way, without specialist knowledge.

Accomplishments that we're proud of

We worked well as a team and we were able to delegate tasks. The different languages we developed in. The overall app.

What we learned

Lots of different languages - JS, Python, R and HTML. How to work well as a team.

What's next for traffiK

Success, shining lights and life-long careers for all of us ;)

Share this project: