How to choose the right graph can make or break a presentation. Sometimes, the choice of a graph is a no brainer, and only one graph type will work. Other times it’s a little more challenging working out how to best present your data.
The decision is made even more complicated now that there is so much data available. The primary reason for a graph is to simplify the findings contained in your data. You want to make sure that message is loud and clear.
So, what are the critical factors you need to take into account when considering how to choose the right graph? The choice of the graph is dependent on a few factors:
- How many variables do you want to present
- Are you presenting time-series data or cross-sectional data
- What is the key point you are trying to get across to your audience
To help make the decision easier, you can think of presenting your data in one of four ways:
1. Comparison of your data
Showing comparisons of your data is something you are probably very familiar with. It’s undoubtedly the easiest method to understand, and we see comparison graphs everywhere. When comparing data, you would generally be trying to show one of the following:
- A rank of several categories in order (i.e., country populations)
- A pattern across a series of data points
The most common graph types for data comparisons would include:
Others could include heat maps, tornadoes, and box & whisker plots.
2. Composition of your data
A comparison chart helps us see each of the data points in the context of the whole data. These graphs help us ‘spot the difference’ in our data and highlight the differences across all data points.
Typical graph types for composition data include:
| Stacked Column or bar
You could also be tempted to use a pie chart, but resist the urge! Read up on Why you need to stop using pie charts.
3. Relationships in your data
Plotting relationships on a graph can help you quickly identify observable trends or correlations. Relationship graphs can also highlight outliers in the data and clusters of data points that might not be apparent in the raw data.
Common graphs for showing relationships include:
A Sankey or polar area chart can effectively communicate the relationship in your data in certain cases.
4. Distribution of your data
Distribution graphs combine the value of comparison and composition charts to present patterns rather than the actual data. The most common distribution graphs are those that show us population across areas, age of a group of people, or how wealth is spread across a population.
Common graphs for showing distributions within your data include:
Choropleths and box & whisker plots are also great options when you have geographical or time series data.
Beyond the basics
There are more elaborate and exotic graphs that you could use, and if you have the right data and the right reason, you should use them. Remember, though, the primary reason for a graph is to simplify the message that’s contained within a great deal of data. You want to make sure that message is loud and clear.
To help work out which graph to use, download our guide on choosing the right chart.