Before you can visualise data, you must understand its nature. Not all data is created equal, and forcing it into the wrong type of chart is like trying to tell a story in a language your audience doesn't speak. It's not just ineffective; it can be misleading. The first step is always to classify your variables.

Data variables generally fall into four categories, which we can group into two pairs: categorical (nominal and ordinal) and numerical (discrete and continuous).

Categorising your data is the foundational step. If you know you have ordinal data, for example, you immediately know that the order of the items on your chart's axis matters. If you have nominal data, you could shuffle the order without losing any meaning.

For categorical data, your goal is to show comparisons between distinct groups. Bar charts excel at this. The length of each bar represents a value, such as count or frequency, making it easy to compare categories at a glance. For nominal data like 'Favourite Pet', the order of the bars (Dog, Cat, Fish) is arbitrary. You could reorder them alphabetically or by size and the chart's message would remain intact.

Ordinal data, however, has a built-in sequence. When visualising customer satisfaction ratings (Very Dissatisfied, Dissatisfied, Neutral, Satisfied, Very Satisfied), you must preserve that order. Placing 'Satisfied' next to 'Very Dissatisfied' would create a confusing and illogical picture. The progression along the axis adds a layer of meaning that isn't present with nominal data.

When we move to numerical data, the visual tools change. While you could use a bar chart for discrete data (e.g., number of children per household), a line graph is a poor choice unless there's a time component. Connecting discrete points with a line implies a continuity that doesn't exist. You can't have 2.5 children.

Continuous data is where line graphs truly shine. They are perfect for showing trends over time or a continuous range because the line itself signifies the connection between data points. A graph of temperature over a 24-hour period makes sense as a line because the temperature exists at every single moment between the recorded points. The line is a truthful representation of the data's nature.

Using a line graph for categorical data is a common mistake. If you plot sales figures for 'Apples', 'Bananas', and 'Oranges' and connect them with a line, you're visually implying a relationship or trend between them that simply isn't there. What does the line between 'Apples' and 'Bananas' even mean? This choice increases the viewer's cognitive load because they must work to override the visual suggestion of a connection. A bar chart, with its distinct and separate bars, avoids this confusion entirely.

At the heart of choosing the right chart is the concept of [{_]. This is the process of mapping data values to visual properties. These properties, sometimes called visual variables, include length, position, shape, size, and colour. The effectiveness of a graph depends on how well its visual encoding matches the data's characteristics.

A bar chart, for instance, encodes a single numerical value as length. A scatter plot encodes two numerical values using position on the x and y axes. A heat map uses colour intensity to represent a value.

The goal is to select an encoding that our brains can decode accurately and efficiently. Research has shown that humans are very good at judging length and position along a common scale, which is why bar charts and line graphs are so effective. We are less precise at judging area (as in bubble charts) and angle (as in pie charts), which is why those chart types should be used with caution.

Choosing the right visualisation is not just about aesthetics. It is about selecting the visual encoding that most truthfully and clearly represents the underlying structure and relationships within your data. Get the mapping right, and the story in your data will tell itself.

Time to check your understanding of these core concepts.

By correctly identifying your data type, you can choose a chart that communicates your findings clearly and honestly.