What is Data Storytelling?

By Storylabs

* Last updated on 9 December 2020

The emphasis on the need for Data Storytelling is being propagated widely, but the term itself means different things to different people. Concepts like the '3 act structure' are useful as a high-level guideline for all stories, but lack a prescriptive detail when it comes to data stories in particular. This work is an attempt by Gramener to bridge that gap.

What is a Data Story?

Data stories are insights that are explained in a particular sequence.

What do BI dashboards lack?

BI dashboards lack two things;

  1. Explanation - Users don't know what to infer. They don't know unless they are told.
  2. Sequencing - Dashboards leave the reading sequence to users, leaving it to them to put together the jigsaw piece and make sense out of it.

How could we introduce explanation?

Explanations are delivered using 4 data-driven narratives;

  1. Context
  2. Summaries
  3. Annotations
  4. Recommendations or take-aways

We can make explanations repeatable through a gallery of templates. Explanations can be linked to data like how charts are linked to data.

We need to explicitly explain insights to make our dashboards communicate better.

Insights are explained mostly using one or more of these structures.

  • Degrees of comparison
  • Up or Down
  • Percentages, Fractions & Ratios
  • Then vs Now
  • X & Y are related
  • Question & Answer
  • Assumption vs Fact
  • Claim -> Fact 1 -> Fact n -> Close
  • Recommendation -> Insight 1 -> Insight n

1. Degrees of comparison

The simplest & most common explanation structure is the 'degrees of comparison' pattern. Insights that talk about the highest, lowest, most, least, best, worst etc belong to this category.

Example: Riverside County has the highest % of adults complaining of insufficient sleep in California.

2. Up or Down

Yet another common pattern is 'up or down'. How has an indicator performed? Has it gone up or down? Has the score improved or declined?

Example: Coronavirus cases around the U.S. are surging - and so are hospitalizations.

3. Percentages, Fractions & Ratios

Readers understand the relationship between numbers better when they are expressed in terms of percentages, fractions & ratios.

Example: Four out of ten who took loans say they will not be able to repay them.

4. Then vs Now

There are times when there's the necessity to call out a major difference between two specific time intervals. That's when we use the Then vs Now structure.

Example: The operational cost of cotton in Punjab was Rs 809 per hectare in 2003, but increased 5X to Rs 5,754 per hectare in 2015!

5. X & Y are related

Two or more factors could be related.

Example: Countries with female leaders tend to have lower Covid-19 death rates and better economic performance.

6. Question & Answer

The Did-you-know or trivia pattern makes people think. It is a useful technique to allow people to self-assess their understanding of information.

Example: Do you know which state in India has the largest cultivation area of cotton? It's Maharashtra. 36% of cotton cultivation area in India is from Maharashtra.

7. Assumption vs Fact

Data many a time disproves popular opinion.

Example: Contrary to popular belief, lung cancer is on the rise among young women.

8. Claim -> Fact 1 -> Fact n -> Close

In this structure we start by asking a question or making a claim. We follow that up by providing evidence. We close by tying it back to the initial question or claim.


Skeptics of manmade climate change offer various natural causes to explain why the earth has warmed 1.4 degrees fahrenheit since 1880. But can these account for the planet's rising temperature?

Is it the earth's orbit? - negligible.

Is it the Sun? - No.

Is it volcanoes? - No.


It really is greenhouse gases. Manmade factors are to blame.

9. Recommendation -> Insight 1 -> Insight n

Certain audiences require instruction upfront. In such cases start off with a recommendation and then get into details.


Increase staffing by 2 persons on Friday & Monday mornings to make sure ZDH part shipments reach on time.

  • Product shipments get delayed mostly on Friday & Monday mornings.
  • The ZDH product category is impacted because of the delays.
  • ZDH part shipments are the most delayed

Why is sequencing important?

Sequences are where the author guides the user to consume the message.

Dashboards leave the reading sequence to users. This makes users take efforts to put the jigsaw piece together to make meaning.

We suggest well crafted linear & non-linear structures to make consumption of insights easier.

Linear sequence structures


The stepper sequence allows the reader to consume one piece of information at a time. The subsequent information is revealed only at the click of a button.

Numbered Tiles

Tiles reveal the overall summary on the first view. More information is revealed at the click of each tile.

Numbered Tabs

Tabs like steppers reveal one piece of information at a time.


Panels (max of 3 or 4), like in short comic strips are useful for the Question & Answer explanation structure.


This interactive technique has a static layer & a scrollable overlay. The visual elements in the static layer transition as the overlay scrolls.


Carousels allow readers to view one piece of information at a time.


This technique is used for maps and aerial photographs. The swipe or click & drag action on a screen allows the reader to view the before and after states of a place.


Videos are a useful linear format to narrate visual stories supported by subtitles or audio narration.

Non-Linear sequence structures

Non-linear sequences allow readers to make choices.

  • Chat interfaces
  • Mind maps
  • Decision trees
  • Filters
  • Sliders
  • Layouts

Putting it together

Explanation & sequence structures come together to make meaningful data stories.

The above example uses the following ingredients to build the data story;

  • Explanation structure: Recommendation -> Insight 1 -> Insight n
  • Sequence structure: Stepper
  • A comic character as narrator (optional)

Explanation & sequence structures give analysts a decent start when faced with the task of creating a data story. Templates created using these structures will save time and make semi-automation a possibility.

If you've found this helpful, please share your suggestions or feedback with us at


  • Bach, B., Wang, Z., Farinella, M., Murray-Rust, D., & Henry Riche, N. (2018, April). Design patterns for data comics. In Proceedings of the 2018 chi conference on human factors in computing systems (pp. 1-12).
  • Feigenbaum, A., & Alamalhodaei, A. (2020), The Data Storytelling Workbook, Routledge.
  • Kosara, R. (2017). An argument structure for data stories. Short Paper. Proceedings of the Eurographics.IEEE VGTC Symposium on Visualization (EuroVis).
  • Nussbaumer Knaflic, C. (2015), Storytelling with Data. A Data Visualization Guide for Business Professionals, Wiley.