Jocellyn's Blog

Data visualizations by hand (Visu homework I.)

I got an interesting assignment at my University this week - to draw a variety of data visualisations of one boring dataset. We should think about the possible representations and about what kind of chart would fit the best to be A) visually pleasing and at the same time B) informative and clear.

It was not a hard exercise but it got me think about multiple stories that can be told with the same dataset.

This was the dataset: (if you are interested, it is the volume of immigrants comming to the Czech republic) data

And this was my drawings:


As you can see, I started with the dummy representations - a pie chart with a ratio between immigrant from EU versus those from outside of EU - in general across all the years. This may look silly but it is a perfcetly valid representation which is telling us something very clearly. There were many more migrant from EU!

Another typical chart is a bar chart with all the years on the X axe and total immigrants regardless of the country on the Y axe. This is in turn showing absolute numbers and it is focusing the time as the interesting factor. We can see how the trend went year by year.

These two are classic examples of visualisations focused either on some kind of ratio (pie chart is best for that) or on absolute volume (the bar chart where the hight of the bar represents the number of immigrants).

Another example of focusing on absolute volume is the colored “world map” (yes, I know the real world doesn’t look like that, exuse my poor drawing). It is the only chart in my selection which is showing us something about total numbers of immigrant but based on countries, not based on years (we could have tell the same story with a bar chart where the countries would been on the X axes, but with so many countries I thought it would be unreadable.)

You can also focus on the “gestalt”, on the shape of the trend. I tried to do that with the simple line chart (with two lines, one for EU and one for non-EU), where we can clearly see that for both categories there was a spike (around 2008) and then a small dropdown (around 2010) and since then, it only went up. We could have missed this trend only from the table.

In summary, there are dozens of another visualisations which I did not draw and which would be perfectly good for this kind of data. The important thing is to know what you are trying to tell. What story are you trying to represent. Do we want to know how much more Asiats are migrating than our neighbours? Lets do a pie chart with the ratio. Do we want to see how is different the trend across the years between some categories? Lets use line chart where we can put multiple lines next to each other. Do we want to see where are the regions with highest absolute numbers of immigrants? Well, lets draw a map (which is definitely the funest thing to draw :)!


comments powered by Disqus