The world’s leading publication for data science, AI, and ML professionals.

This Is How You Should Be Visualizing Your Data

10 Examples to Guide Your Analysis

Photo by Chris Leipelt on Unsplash
Photo by Chris Leipelt on Unsplash

Introduction

Have you ever been told you have been doing something wrong the entire time? The way you train, eat, speak, sleep, react, think… Every discipline has the way we do it and then some Ted Talk of how we should be doing it.

Now I am not going to be doing a Ted Talk on data visualisation, but I am going to provide you with 10 tips on knowing which visualisation to use for the type of data you want to show.

Before I begin, there are three points to keep in mind when creating data visualisations:

  1. Keep it according to the needs of your users – It is not about what you want to see.
  2. Keep it simple – This makes it easier to understand.
  3. Keep it relevant – Your users must be able to see and understand the main findings within 5 seconds.

Given that I am preaching that less is more, let me keep this short and informative.

1. Line Chart

Line charts are great for showing how things change over time.

Examples:

  • Plotting sales over time to identify seasonality or gauge how sales have been performing.
  • Retention rates – At what point in time are online customers churning?
Source by Author
Source by Author

2. Column Charts

For comparing performance across categories.

Much like the line graph, one can visualize the sales over time, but for separate companies, brands, products etc.

Example: Growth in market share of the current period vs previous period.

Source by Author
Source by Author

3. Bar Charts

When you flip the column chart on its side – you get a bar chart.

Where a column chart is more suited for showing the values of different categories over time, a bar chart is more suited for comparing different values for different categories.

For example, below you can see that the most orders placed are for Office Supplies, however, this gives the company <₤10 profit.

Source by Author
Source by Author

4. Pie Charts

When you need to show percentages per category at a specific point in time.

This is a visual way to show the differences in size. Pie charts are mainly for showing high level Key Performance Indicators (KPIs) and should not compare more than 8 categories – just like a pizza should not be sliced into more than 8 pieces…

Photo by mahyar motebassem on Unsplash
Photo by mahyar motebassem on Unsplash

Example:

  • Market share of your brand vs competitors.
  • Survey data – What % of people said yes vs no?
  • Percentage conversions coming from each marketing channel.

Do create pie charts like this:

Source by Author
Source by Author

Do not, under any circumstance, create pie charts like this:

Source by Author
Source by Author

5. Area Charts

Source by Author
Source by Author

An area chart is a more visually dynamic version of the line chart.

Where a line graph shows a time series trend, an area chart can show multiple trends as well as their comparative sizes over time.

One would use an area chart if there is a clear difference in the size of the various trends you are plotting. For example, in the chart above, one can clearly see that Technology brings in more sales than Furniture.

6. Highlight Tables

Source by Author
Source by Author
Source by Author
Source by Author

This is my personal favourite. So much so, that I have written two tutorials on creating these in Tableau ([here](https://medium.com/analytics-vidhya/creating-your-first-dynamic-tableau-dashboard-c373adacd026) and here).

If you need a visual way to summarize several measures in a large dataset, then highlight tables are ideal. You can guide the viewers’ eyes with colour and symbols, making it easier to digest.

7. Scatter Plots

Source by Author
Source by Author

Scatter plots work well when you need to visualize two different variables to see if there is any correlation between them. In the example above, one can see that there is a positive relationship between sales and discount. The colour of the circles helps differentiate the category that each data point falls under, and the size indicates the profit.

Only use a scatter plot if there is a (negative or positive) correlation between the two variables. Make sure not to overplot:

Source by Author
Source by Author

Scatter plot examples:

  • Experience vs wage.
  • Hours spent studying vs test mark.
  • Ad channels spend vs conversions.

8. Map Charts

A map chart is used to visualize geographical data. You can overlay circles on the geographical data (using the same design principles as a scatter plot – colour and size) or you can transform your map into a heat map, where you only make use colour:

Source by Author
Source by Author
Source by Author
Source by Author

Examples:

  • Which countries are most profitable?
  • Where are the most downloads for your app coming from?

One needs geographical coordinates to make a good area chart. You can also use country or city, but the more granular your level of detail is, the better the area chart.

9. Treemaps

Source by Author
Source by Author
Source by Author
Source by Author

Treemaps are used when you have hierarchical data, where you make use of colour, the size of the rectangles and labels to paint your picture.

Examples:

  • Visualising the 10 major symptoms of Covid.
  • What are the most profitable products and their respective shipping modes?
  • When you want to show the relative sizes of various product categories as well as indicating which are more profitable.
  • Votes for a specific political party by state.

10. Numeric Indicators

Numeric indicators give a high-level overview of the direction in which your main KPIs are going. For example, certain stakeholders are only interested in seeing if the company/product/segment is profitable or not:

Source by Author
Source by Author

Alternatively, certain stakeholders may be more interested in the details of the data. However, by including the summary numeric indicator, you can get an idea of the bigger picture before diving into the details:

Source by Author
Source by Author

Conclusion

Photo by Customerbox on Unsplash
Photo by Customerbox on Unsplash

The creativity in data visualisation is endless and this article only touched on the basics. Feel free to reach out for additional resources.

Please consider using my referral link if you would like to become a Medium member and supporter. Your membership fee directly supports all writers you read. You’ll also get full access to every story on Medium.


Related Articles