Vis-Rec is a tool that visualises how the service recovery after an incident at a station or on a line, effects the connecting lines and stations. Using the historical data it shows how the service recovery over the day has brought the system back to its normal running status by highlighting delays using a RAG status. This is a vast improvement to the static graph that managers currently receive after an service recovery has taken place.

We can then use this visualisation to easily compare it to similar incidents that have occurred at the same station. It will clearly show which was the more successful strategy and will therefore give experts in the domain of service recovery, to analyse what decisions made it more effective hence in the future look to standardising the approach of service recovery to specific incidents. When an incident occurs in real time train running controllers can use the best practice visualisation to react to the incident.

This was built by first understanding what data we would need to aquire from darwin to establish which stations connect to the line that has been affected. We also needed to compute how much of a delay each station has at a given time step. For our PoC of Birmigham New Street & surrounding areas, that's 30 stations statuses over a whole day showing the delay data every 30 minutes. Once established what data we would use, we used Google BigQuery to read all the data we needed for our vizualisation. We used React & Leaflet to vizualise the data that we build up from Darwin & BigQuery.

The main challenge was understanding the data. having to cross reference very different data that was in different formats. This was extremely time consuming and we made some assumptions from a certain set of data only to realise that it wouldn't work with the other data. We also changed direction in what solution we would create at 16:00 on the saturday which meant we had used up a lot of time developing the wrong solution.

As a team we had a great brain storming session when we realised that our original idea wouldn't work. We were pragmatic but also supportive of each others ideas. Once we decided on our new approach we assigned roles efficiently and began work quickly. we worked to our strengths and supported each other as much as we could doing regular updates on what we were working on, how it was progressing and where we could do more.

Personally I learnt that you need to ensure you are always supporting the team, if you can't do a hard technical task then it's your job to take the mundane jobs that will free up the people with the technical skills. I learnt a lot about data sets, the importance of understanding them to extract the correct information and the importance of really understanding how each set of data is truly effecting every thing else and to never make assumptions.

Vis-Rec has the cability of digitilizing and improving how quickly you can establish if a recovery after an incident went well, and from learning what was a good service recovery by working with analyst this comparative visulaisation can enable the production of a standardised process to reacting to an incident. This will take the pressure of the individuals who are working that day and will allow for improvement in the service recovery hugely impacting passengers.

Share this project: