How to Choose the Right Visual Representation
This is some text inside of a div block.
Types of Data Visualization
The first critical step in data visualization is choosing the appropriate type of visual representation that fits your data. To do this, you’ll need to be familiar with the options. So let’s break down the main types of data visualizations:
Charts
Charts are a way of representing large datasets in a condensed, easily digestible way. The term “chart” is a broad category that encompasses graphs, tables, diagrams, and maps. Depending on conventions in your field, “chart” and “graph” may be used interchangeably.
Graphs
Graphs are the most common type of data visualization used by scientists and researchers. This is because graphs offer a simple way to represent comparisons between groups of data across two or more variables. To convey these differences, graphs use points, lines, segments, curves, or areas. We’ll expand on how to choose the right type of graph for your data in the next section.
Create clear, beautiful visualizations of your research data with BioRender Graph. Learn more today.
Tables
Tables are used to organize data into rows and columns. Although tables are not the most visually compelling way to display data, they are incredibly useful when individual values are important, as opposed to comparisons between aggregated data.
One way to make tables more digestible is to replace text with visual elements. For example, when heat maps are organized as tables, they use color gradients to visually depict comparisons between individual values. How the data relates to one another is more easily understood at a glance.
Geospatial
Geospatial charts are used to superimpose quantitative data on top of maps. These are ideal if a key element of your data is geographic location, which can be represented with shapes, colors, or numbers along latitudinal and longitudinal coordinates.
Infographic
Infographics combine illustrations with descriptive text and may include data charts. The primary purpose of infographics is to provide a high-level overview of a topic in a very visually appealing way. Because of this, summary data is often included, while detailed analytics are typically excluded.
Dashboard
Dashboards are a consolidated collection of related metrics and insights derived from multiple data sources. They can be static but are most useful in an interactive form that allows for data filtering and real-time updates. For example, regularly updated COVID dashboards helped millions of people track relevant pandemic statistics online.
In research papers, multi-panel figures with related charts often function as static dashboards to provide a more complete analytical picture using multiple data sources.
How to Choose the Right Graph
For scientists and researchers, graphs are by far the most commonly used type of data visualization. Choosing the right graph is the next critical step in clearly, accurately, and efficiently communicating your data.
The right graph should:
- Clearly show how values relate to one another
- Represent quantities accurately
- Make it easy to compare quantities
- Offer a ranked order of values
- Make it obvious how the audience should use the information
We can categorize graphs into four main types based on data they best represent.
- Comparison: For comparing different groups or categories, highlighting similarities and differences (highest/lowest points in data, increases/decreases over time, etc.).
- Example: A column chart comparing the effectiveness of different drug treatments in terms of patient recovery rates
- Relationship: To display the correlation or connection between two or more variables, showing how changes in one variable relate to changes in another.
- Example: A scatter plot illustrating the relationship between the dosage of a medication and the corresponding reduction in symptoms over time
- Distribution: To represent the spread and pattern of a dataset and draw attention to the frequency and variability of values
- Example: A histogram depicting the distribution of blood pressure levels in a population to understand the prevalence of hypertension
- Composition: For showing the parts of a whole, which helps visualize the contribution of each component to the total
- Example: A pie chart illustrating the proportion of different types of cancer diagnoses within a specific demographic, indicating the relative prevalence of each type.
Below is a non-exhaustive guide to selecting the right graph based on your data. Note that these categories are not always mutually exclusive.
For example, distributions are sometimes considered to be a subcategory of relationships, and some types of comparisons (like a table with embedded charts) can also be classified as relationships. Here is a commonly used guide to simplify decision-making (it is not meant to provide strict definitions).
BioRender Graph Tip
BioRender Graph makes it easy for you to choose the right graph for your analysis. When adding a new data set, you’ll be prompted to select the type of analysis you want to perform. Based on your selection, a preview of compatible graphs will be displayed.
Read the next chapter | Return to the table of contents
References
- Data Visualization: a practical guide to producing effective visualizations for research communication by Rebecca Wolfe (2014. RESYST Consortium)
- Tableau article: What Is Data Visualization? Definition, Examples, And Learning Resources
- Tableau article: A Guide to Charts
- Tableau article: A Guide to Tables
- Tableau article: A Guide To Geospatial Visualizations
- Datylon article by Dieuwertje van Dijk: 9 types of data visualization
