… and presenting data

Today I worked more on the slides and material for Data Engineering, on visualising data. I spent some time making a cool illustration of how not to make a graph:

A bad graph

Hopefully some of your eyes are bleeding just looking at that graph.

Tonight we’re playing online board games, and concentrating on Castles of Burgundy, a new game that one of your group has been urging us to try for some time.

New content today:

One thought on “… and presenting data”

  1. Let’s see:

    -Y axis does not originate at zero
    -Y axis labels are not scaled correctly. That is, the numeric values represented by equidistant ticks are not proportional. Two of the intervals are 50, but one is 100 for some reason.
    -Y axis itself needs a more descriptive label, say “Drink sales per day per store” or “Drink sales, per barrista per day, in Australian dollars”
    -Rotation of the 3D graph makes it hard to visually judge height of graphical “bars” in bar chart
    -Use of graphical “bars” is misleading because we are supposed to judge by height, but there is a human tendency to judge by volume, so something twice as tall might look four times as “big”
    -X axis itself is marked as “size”, which means nothing without context. A better label would be, “Drink size” or “Weight of patron” or “Size of patron’s bank account in Australian dollars” or something.
    -X axis data points not labeled. Assume it’s “Drink size”. Each one should be labeled, say “250ml, 400ml, 750ml”.
    -Graphical background is confusing. Replace with horizontal lines (typically dotted) extending from the y-axis, making contact with the top of each “bar”, to allow the reader to immediately know the numeric value for each “size”. Alternatively, put a label on/in each bar showing the number it represents.

    How did I do, professor?

Leave a Reply

Your email address will not be published. Required fields are marked *