Choosing the right visual for the right data — Every single time.
Bar Charts, Histograms, Pie Charts, Scatterplots, Treemaps, Heatmaps, Dendrograms, Stacked bar charts, and a million other graphs. It takes a pro to decipher which graph would work best within a given use case. It all comes down to 3 major things, that we will go over in this article to get you all loaded for your next data visualization task at work/academics!
- What is the story you want to tell with your visualization? (The Role)
- What is the data you are using and what is the variability? (The Data)
- Does your chart convey helpful information? (The Result)
Also, read (if you like Dataviz):
What is the story you want to tell with your data? (The ROLE)
It is the first and foremost requirement that you understand the story you want to convey with your data. Knowing what your audience expects to see and knowing the information you have to share with them will help you frame a better idea of the story you want to tell.
This can be done in the following different broad categories of stories that you can communicate:
Single value story:
Sometimes your entire story can revolve around one number, so it’s best to always show this number clearly on your visual with elements that can additional information to the number without overdoing it.
For example, the Churn rate in your human resources.
Change over time (Time Series Data):
If you wish to take the same churn rate example and want to portray how it has been changing over the years, it would help to use a line chart, bar chart, scatterplot, sparklines, etc. that can use the timeline in the X-axis and show how things have been moving.
For example, Stocks in a Stock market.
Comparison:
Sometimes your story can revolve around an A/B test where you compare the results of a training program or a change in your current workflow. To show the comparison between two or more things, you can use a side-by-side bar chart or stacked bar chart, or continuous line charts to portray the comparison.
Part to a whole:
More often than not, very basic visuals are all about showing a part-to-whole comparison, where pie charts, etc. are used. However, to avoid sharing information in pie charts, for the same data, a bar chart or an area chart would be better options.
Relationships:
When A and B are two variables that are related to each other, for example, the churn rate of employees and their time in the company/salary, it would be best to show the relationships using a scatterplot. Though, a scatterplot would be the best plot you can use for relationships; heatmaps and bubble charts allow the same outcome.
For example, the relationship between the height of a cylinder and its diameter.
Flow:
When the flow of a certain process or a pipeline is concerned, an alluvial diagram or a funnel chart would be great. Other options for the same like Gantt charts and Dendrograms would be useful too.
For example, A sales funnel chart (as given below)
What is the data you are using and what is the variability? (The DATA)
There are different types of data that you can have in your variables. The important aspect of this is to understand which visual fits which type of data in the best way. For example, using discrete variables (especially categories given in text format) might not be the most appropriate way of using a scatterplot.
Running your data through some data exploration and taking the descriptive statistics will also help you take better decisions in choosing the right type of visualization to emphasize the points you choose.
Every datatype has its visualization to use to amplify how the information is shared with a viewer. Here is a non-exhaustive list of how you can use the data visualizations to suit specific data types.
Note: It is out of the scope of this article to explain each graph in detail — because that would be better if explained in their own articles. You’ll find resources to understand the visuals better and some things to remember, at the end of the article. Do stay tuned and let me know in the comments if you want me to elaborate on the graphs too!
- To make any of the graphs below, try using a data viz tool like Tableau or PowerBI, they’ll help you make the right visual for the right data by choosing for you.
Continuous Variables:
For continuous variables, one could always use the following graphs:
- Histograms
- Line graphs
- Scatterplots
Discrete Variables:
For discrete variables, one could always use the following graphs:
- Bar Charts (simple, stacked, side by side, etc.)
- Density curves
- Box plots
Textual data:
For variables with textual information, one could always use the following graphs (depending on the use case) to communicate insights:
- Word Clouds
- Bubble Charts
- Network charts
Ranked or Ordered (Ordinal) variables:
For ranked variables, one could show the order using the following:
- Compound Bar charts
- Treemaps
Distributions:
For showing distributions clearly, one could always use the following plots:
- Histograms
- KDE Plots
- Rug plots
Note: YOU DON’T HAVE TO USE COMPLICATED VISUALIZATIONS.
The importance of data visualization is to easily communicate the data insights to an audience. Adding layers of unnecessary complication to your visualization will only make it harder for the audience to perceive the data in the way you want them to and will also make it harder for you to communicate the message even if the visual looks top-notch and has all the information.
P.S. That’s the key. It does not have to have ALL the information, just the parts that are RELEVANT.
Does your chart convey helpful information? (The Result)
When you have your data and story taken care of, it’s time for improvement. It’s the best thing in the world if you can ace your visual in the first go and make a dashboard that can be used by the audience. However, most visuals always have a lot of scope for improvement.
Here is where your network comes into the picture.
If you are a student making a visualization for a college presentation or project, it’s best to show it to your peers and take their opinion on it. The same applies to corporates who have a team to show it to, or just about anyone in your organization. Having answers to the following questions can help improve your visualization exponentially:
1. What is the first thing you see when you look at the visual?
2. Is the information clear to understand?
3. Are the colors too bright or the contrast too less?
4. What thoughts are you left with once you’ve assessed the entire visualization?
5. How long did it take you to go through the entire visualization?
Improving your efforts on visualizations is the best way to learn. If you put in 2 hours to make a beautiful visualization with the right information and aesthetics, take another 20 minutes to make sure that you don’t leave anything unchecked.
A review by a peer will also help you see any overlooked details and prepare a concise and clear visualization to share with your audience.
I hope that this article helps you to choose a better visualization and make your presentations simpler with visuals taking the lead. It is a whole other story why people prefer using visualizations today and not numbers in their presentations. Know more about them here, because its remarkable to know the effect that a visualization has on your brain subconsciously:
Let me know in the comments below if you have any other pointers or charts that everyone should look into. Leave a clap and follow to stay in touch with any new articles and to support the blog!
Also check out my previous articles here, if they interest you!
For more such articles, stay tuned with us as we chart out paths on understanding data and coding and demystify other concepts related to Data Science. Please leave a review down in the comments.
Check out my other articles at:
Do connect with me on LinkedIn at — Yash Gupta — if you want to discuss it further! Leave a clap and comment below to support the blog! Follow for more.
