Visualising networks

For a while now I’ve wanted to start exploring different ways of using mapdeck to make eye-catching and unusual visualisations. So hopefully this post is one of many along this theme.

The goal here isn’t necessarily to meaningfully interpret data, but rather to see what’s possible. And if a story jumps out then great, we can use it.


What’s going on here?

We’ve kindly been given an extract from the Victorian Department of Transport’s TrainSUM model for a single, random Thursday (pre-COVID). Symbolix were fortunate to collaborate with the Department to develop TrainSUM. The software combines survey, ticketing and timetable data to enable analysts to estimate train service loads across the Melbourne Metro network.

Visualising the differences between stations at different times of day is challenging because you need to balance spatial and temporal data. It’s difficult to get a sense of the ‘movement’ on the network. That’s what I’m experimenting with here. In the map above I’m showing the number of people entering each station throughout the day.

And yes, they are entering from the sky, but that’s neither here nor there…

The colour and height of each ‘fall’ represent the number of people*, so the higher (and more yellow) the fall, the more people are entering the station.

Is it any good?

Yes. Next?

Ok, why is it good?

A few things I like about it are

  1. The sense of movement of people. I feel like I’m watching them enter the stations.

  2. Having the height represent the number of people, and it changing over time, to me shows a good comparison between how many people each station has to cater for.

  3. If you watch carefully you can see the trains passing through as the number of people entering momentarily goes to zero.

  4. The manipulation of the data I did behind the scenes to make it.

How could you improve it?

I have a few ideas, including showing the actual trains travelling between stations and people leaving the network too., but that will have to wait for another post.

What do you think?

Is this a good way to represent people entering the train network?

*For actual patronage values please refer to DOT