If data visualizations are a central element for understanding data analysis, choosing the right type of graph is even more important. Various options present information differently, each with its advantages and disadvantages.
When visualizing data, graphs are a central point. The first thing to consider before choosing the visual tool is the objective: What story are these data trying to tell? Are you looking for trends? Comparing variables? This is the first premise; the graph should align with the storytelling.
The second point is to be clear about who will see these results. It is different to present data to a general audience compared to a specialized audience or people trained in statistics or data.
The size of the data will also significantly affect the type of graph, as some are not intended for massive data groups. For example, pie charts work better with a small number of datasets; however, if you are using a significant amount of them, scatter plots might make more sense.
There are various types of data, which also conditions the choice. For instance, if you have continuous data, a bar chart may not be the best option, but a line chart might be. Similarly, if the data is categorical, then using a bar chart or a pie chart could be more convenient.
Another point is how different elements of the data relate. Are there variables such as time, size, type? Are they a time series? Is it a distribution?
One of the most common uses for data visualization is to show how a variable’s value changes over time. These charts usually have time on the horizontal axis, moving from left to right, with the values of the variable of interest on the vertical axis. In these cases, bar charts or line charts are usually the most suitable.
Sometimes, the goal is to understand how a whole is composed. In these cases, pie charts, ring charts, or stacked bar charts and area charts can be very useful because the components are highlighted.
If the goal is to compare values between different data groups, bar charts, line charts, dot plots, or grouped bar charts can serve well.
If the objective is to understand the relationship between data features to observe trends and patterns among them, scatter plots are the standard way to show the relationship between two variables. If there are additional variables, bubble charts can represent them well.
It is important to experiment with different types of graphs and see how the data is presented in each case and in relation to the storytelling you want to convey. It is also important to keep in mind that it is not necessary to show everything in a single graph. Often, it is better to keep each graph as simple and clear as possible and, if necessary, use multiple graphs to make comparisons, show trends, and demonstrate relationships between multiple variables.
By Julio Cesar Blanco – August 22, 2022
