The easiest way to explain your data to someone else is actually to visualize it. Data Visualization could be considered as story-telling or how to tell your story to your audiences or as analysis tool to make you understand your data better.
To be able to visualize data, there is a very useful knowledge called ‘Semiology of Graphic’ which many seems to overlook and go to what chart I should use. Semiology of Graphic deals with how we can actually play with different aspects of the chart to reflex the differences in data.
(Image Source: https://www.axismaps.com/guide/general/visual-variables/)
One of the tools that really works well on this Semiology of Graphic is actually Tableau.
I’ve provided some samples for common visualization practices below.
1. Conditional Formatting is a lazy way to quickly understand data. I’d say this is the most overlooked visualization. By just adding conditional formatting, it was very clear which cell is outstanding in good or bad way. It is also very easy to implement just 1-click in excel and you already can get this table.
(Data from Passport: Digital Consumer Index)
2. Line Chart / Bar Chart is good when you want to see the trend or magnitude of each value. It is much easier to see when you visualize it out as Line Chart (on the right) than just table (on the left). Line chart is really good for continuous data, especially changes over time.
However, I’ve one thing that I’d forbid everyone I know to do. Do not use Line Chart for categorical data. The example below show how you cannot put Line Chart for comparison among male / female as the between you doesn’t have ‘ladyboy’. Thus, you should use bar chart in this case instead.
3. Tree Map / Pie Chart is good for plotting ‘part-whole’ relationship or how big is the specific portion is as you can see from below chart for SEA population. Tree Map could also show hierarchical data (such as countries > cities) but Pie Chart can reflect ‘part-whole’ relationship better. Nonetheless, both are not suitable for comparison.
4. Stacked Bar Chart to show composition (which could be combined to 100% or just the composition). The example below from Consumer Barometer showed both comparison and proportion of each.
5. Scatter Plot is your best friend to see 2-dimensional data. The chart below from the data on the left made it clear that Singapore is lowest in both App and Web.
6. Divergent Bar Chart is very suitable for Likert scale where you have both positive and negative answers such as this visualization of Wiki4HE Dataset which shows clear positive answers.
Just to share that these are only sample on how you can visualize the data. There are much more chart and graph types that you can use such as network graph or map visualization and you can be even more creative and utilize more Graphical Semiotics!
Ps. I also found it very useful to go through this checklist for any visualization I’ve created.
At Davoy, we could provide you with Data Visualization consulting and training. Just fill in the contact form and we’ll get back to you for free first consultation!
Interested in transforming your reports into dashboards? Contact us at Davoy.tech via the form below