How to Design Effective Charts with the Help of Statistics

Short statistical theory

To create a graph, we must first have data, which is a collection of numbers that describe certain variables. These variables have specific values, such as gender, age, amount of money, etc. In statistics, there are three types of data:

1. Nominal data, which cannot be ranked. Examples include hair colour, languages, countries, and gender.

2. Ordinal data, which can be ranked in a specific order. Examples include education levels and military ranks.

3. Interval data, which can be ranked and have equal intervals between each value. Examples include age and weight.

Here’s a quick tip on how to understand what kind of data is in front of you (picture 1).

Picture 1: How to figure out your data type

Graphs

The goal of creating a graph is to present complex information in a simple and easy-to-understand way. With an understanding of the different types of data, we can now determine which types of graphs are best suited for each.

1. Numerical data. Pie chart

The pie chart is perfect to visualise the ratio between variable categories. To bake a perfect pie, we take some fresh numerical data. Because numerical categories are not rankable, we can sort them by percent (I prefer descending order so that we read them clockwise) (picture 2).

Picture 2: Pie chart

2. Ordinal data. Simple Bar Chart

If you have rankable data, the best way to present it is a simple bar chart because we can sort bars in a meaningful order. Here’s how we read the info then: top to bottom — education levels, left to right — the number of objects (picture 3).

Picture 3: Simple Bar Chart

3. Interval data. Histogram

When we have a long range of data with equal intervals, the best way to represent it is the histogram. A histogram looks like a bar chart, but instead of values on the scale, there are ranges (picture 4: 0–10, 20–30, etc.). Merging continuous data into ranges improves our perception of it.

Picture 4: Histogram

There are many different ways to present your data in a chart. You can compare the number of cases, their percent of the whole sample. The goal here is to demonstrate the result of the research as clear as possible.

Breaking Charts Stereotypes

Pie charts have been overused and may have a bad reputation, but when used thoughtfully, they can be effective. Consider alternative styling if you have negative feelings towards pie charts (picture 5).

Picture 5: A few Styles of Pie Charts

Tables have a reputation for being difficult to read, but when designed well, they can effectively present large amounts of data. Try giving it a chance next time when you need to be precise with numbers.

Sometimes, the best way to present information is through text, without a graph. If only a few numbers need to be shared, simply write them out (picture 6).

Picture 6: Usage of Text

Summary

By understanding the basics of statistics, we can create more effective and visually appealing graphs to present trends and correlations in data. Keep in mind that certain types of data representation may have a bad reputation due to overuse and misuse.

___

Source of data for graphs

Graphs were generated with IBM SPSS Statistics Viewer and edited with Figma.