Story image

Data scientists shouldn’t tell data stories alone - Tableau

11 Oct 2019
Twitter
Facebook

Article by Tableau Software APAC technology evangelist Mac Bryla

Most executives today talk about data as one of their most valuable assets.

Some say data has similar properties to oil - that is, it’s an unrefined, raw resource that can be processed to create a high-value product.

But that might not be the best way to describe its value.

Oil is well understood - we know where to find it, how to extract it, refine it and convert it into fuels and other by-products, and deliver it to users.

Data is nothing like that.

The whole process from sourcing data to producing a finished product is not nearly as well understood.

We don't even know how to find it in many cases, and when we do find it, we tend to stockpile it, thinking maybe one day we'll figure out what to do with it.

We have to clean it up and transform it into a format that we can understand, and then we can start analysing it.

Through that process, we might still arrive at a point that there's nothing of interest or value produced.

If we do find insights in the data, we need a way to communicate those insights effectively with others.

We need to tell stories about the data and we need to tell the stories behind the numbers.

Why? Because storytelling works as a way of sharing insights and getting your point across.

The work of American academic Chip Heath at Stanford University is often cited as proof of the power of storytelling in getting people to remember data points.

Heath had students make short presentations, before distracting them with a short video.

He then asked them to recall the presentation content:

“In the average one-minute speech, the typical student use 2.5 statistics. Only one student in ten tells a story. Those are the speaking statistics. The ‘remembering’ statistics, on the other hand, are almost a mirror image: When students are asked to recall the speeches, 63 percent remember the stories. Only five percent remember any individual statistic.”

The data storyteller

Some industry observers argue the need to tell stories about data can be met by a new type of job role: the “data storyteller”.

The data storyteller builds on the rise of the corporate storyteller over the past decade (or more). There are almost 700 Australians with the job title of ‘chief storyteller’, according to LinkedIn.

Make that ‘storyteller’, and the number climbs to about 3000 people.

Storytelling today is often linked closely to brand and marketing activity, but its roots are firmly in making sense of data.

Database seller Dun & Bradstreet, for example, sees data storytelling as a critical skill for its data scientists, helping them communicate the meaning of data to people that do not have an analytics background.

Data scientists are doing some amazing things, but may struggle to communicate those ideas with IT and business people.

Their work is interesting but complex; it can be difficult to summarise without losing some of the nuance.

A role for technology

Too often, data scientists over-rely on technology to tell their stories.

Technology is a crucial element of storytelling.

A platform like Tableau can help to visualise data in a way that supports and illustrates the stories that the data scientists want to tell.

But good storytelling requires more than just technology.

Good stories are simple, expressive, honest and real.

Telling engaging stories is an interpersonal 'soft skill' that needs specific focus and development.

We should all tell stories

Storytelling is being pitched to solve a translation problem - a way to make the work of data scientists memorable and relevant to the rest of the business.

While stories might appear to make data science more accessible, they do not address the root of the problem.

Data scientists need to speak the language of the business, but equally, the business needs to speak the language of data.

Data is not just for the data scientists.

It's for everyone. We all need to be fluent in data to get the best from it.

It does not make sense that only some people achieve fluency.

When there is a common level of data literacy in an organisation, everybody can participate in like-minded conversations.

More importantly, everyone can tell their own stories.

So rather than invest in storytelling skills for a select few, everyone should be given the skills and foundational knowledge to speak the language of data.

Only then will the true value of data be realised.