Create kick-ass data visualisations using these four steps (with examples) #23

Share on linkedin
Share on twitter
Share on whatsapp
Share on facebook
Share on reddit
Share on email
Share on pinterest

Charts, infographics, statistics, mean, median and bell curves! There is a lot of data and even more ways of presenting it. Don’t get lost in the trees by ignoring the woods.

 

Telling a story using data has never been easy. It requires one to process, organise and coherently present information.

 

More often than not, the entire presentation comes crumbling down at the first question from a knowledgeable person in the audience. The scenario also occurs when the “so-what?” of the story has not been answered.

 

When you have been working on a particular data story for too long, you may be missing crucial information that you cannot spot.

 

It is essential to follow these key steps to get your Data Visualisation to stick in your audience’s mind.

In this week’s post, I will be sharing how I have used these four steps to improve my storytelling.

#1 Define your so-what?

 

#2 Use basic but effective visuals

 

#3 Use a consistent colour palette

 

#4 Create your story

#1 Define your so-what?

The story you are about to narrate to your audience must have a purpose. You want your audience to be engaged and care about whatever it is you are presenting.

 

To do this, you must avoid them from thinking “so-what?” at the climax of your story.

 

The story must be tailored to your audience at the right level with the “so-whats?” answered.

 

Let’s take the following table as an example:

The analysis includes deaths involving COVID-19 that occurred between 2 March and 10 April 2020, registered by 17 April.

 

This table shows the count of deaths based on ethnicity. Straight away, a number jumps out at you from the first numerical row. However, there is no context provided for this table. The numbers are also not ranked to see which community has been impacted from pure counts.

 

Purely from a presentation perspective, this table is merely a dump of information.

 

Let’s now provide some context to it.

 

We have arranged the table in descending order and added percentage points to provide context. It is now clear that from the nearly c13K deaths registered, almost 84% were people from White ethnic background and 6% from Black.

 

However, providing absolute values with no comparison to the size of the population is another way of poorly representing the data.

....providing absolute values with no comparison to the size of the population is another way of poorly representing the data....💭

#2 Use basic but effective visuals

Now that you have some data and hopefully the question you are trying to answer using that data, your next step should be representing this data.

 

Let’s continue from our example above. Representing the table format in the pie chart below is making the data unclear.

 

 

Yes, I can tell the most significant piece of the pie is represented by the White community; however, I cannot determine the difference in sizes between Indian, Bangladeshi, and Pakistani and whether the comparison of these three communities is against the Black community.

 

Using incorrect visuals to represent data is a critical issue. Humans struggle to interpret angles, and as such Pie chart is a poor way to describe data with many different categories.

 

Let’s take another example:

Can you tell what representation the blue part of the pie has?

 

It isn’t easy to spot; it seems less than 50% and at a stretch even 40%. But nothing more specific.

 

Look at the same pie chart turned anti-clockwise.

 

You can easily spot that the blue area is 25%.

 

This representation further proves how we humans struggle to interpret curves and angles with exact measurement.

 

Stephen Few writes beautifully about the pie charts here.

 

So, what should you use?

 

The bar chart is an under-rated visual for representing absolute data. It solves almost all your problems, all the time. It clearly articulates even minor differences without needing to add more context to the visual diagram.

#3 Use a consistent colour palette

The Colour palette is essential as the colour could be loud enough to grab attention or dull enough to make the audience focus on other visual areas.

 

It would be ideal not to use colour and shape to simultaneously grab the audience’s attention, as this will confuse them.

 

You probably didn’t notice the grey coloured bars in the previous section. However, now that I have added a colour contrast to this chart, your eyes are drawn by it. The colour contract is adding absolutely zero value in this visual.

 

So, use the right colour palette but do not use colour unnecessarily.

 

There are many online resources that you can use to find the perfect colour combination.

 

I tend to use https://coolors.co/ as it helps me find visually pleasing colour combinations. The basics also apply here, do not use red, green, and amber.

 

They usually have connotations applied in corporate environments.

#4 Create your story

Now that I have understood the above, it is now time for me to craft my story. Your story must have a beginning, a middle and an end (climax).

 

Depending on how the story is communicated, you could use motion visuals/animations also. Of course, you can’t do that on a PDF that you may be distributing to the executive committee.

 

At this point, you may well find that the information you obtained in the first three steps were insufficient, and you need to find more data to complement it.

 

Introduction — We analysed the number of deaths relating to Covid-19 in England and Wales for the six weeks between early March and mid-April.

 

 

Beginning of your story — We discovered that overall, there were 12,805 deaths. This number is more than what was reported six weeks before that. There seems to be an upward trend in deaths relating to Covid-19.

 


Some data has been mocked up for illustration purposes

Middle of your story — We break this down further to understand the impacted ethnicities.

 

We standardise the above graph with the help of the population size of each ethnicity.

 

We then report death data based on per 100K population. i.e., how many deaths from ethnicity based on 100K of their population.

 

Suddenly, the story has been turned upside down with this discovery. It is evident from the graph above that the Black community are victims of Covid deaths at more than twice the rate as the White community.

 

End of your story — This is your climax. The audience has been engaging with you; you now need to rest your case with your findings. In a corporate environment, we agree on the next steps, call to action etc.

 

The size of the ball represents the likelihood of these minority community members falling victim to Covid-19 compared to the White community. We now need to work out an action plan to shield members of this community from the spread of Covid, and here is our action plan XYZ.

Conclusion

These are the core basics of data visualisation that you ought to perfect to help deal with 90% + scenarios.

 

Answer your “so what?” get the right simple visual, use an appealing colour palette and have a good storyboard for your results.

 

Do you agree with what I’ve said above? What are your thoughts? Feel free to reach out to me via my email at [email protected] if you have some feedback or want to say hello!

 

If you’re still reading this, I hope you’ve found some value in this blog post.

 

Check out my other post below.

Share this post

Share on linkedin
Share on twitter
Share on whatsapp
Share on facebook
Share on reddit
Share on email
Share on pinterest
Hanzala Qureshi

Hanzala Qureshi

I’m a digital consultant at a leading consultancy firm. I mostly spend my life working on complex data projects. On this website I document my journey in consulting and thoughts on data & emerging technologies.

keep reading

Leave a Comment

Your email address will not be published. Required fields are marked *

This website uses cookies to ensure you get the best experience.