What is Data Storytelling?

Posted on December 31, 2018 in
3 min read

It's a controversial term but, briefly, it can be defined as a way to convey data-driven information exploiting some paradigms of storytelling.

The data storytelling ultimate goal is all about lowering the bar to read and understand complex information.

It's particularly relevant when dealing with regular people, in the sense of non-experts in data field. There is a nice and very active community made by academics and practitioners debating around the field but, for some reasons, regular folks are not really included.

While data literacy is increasing in general, non-expert people still need some help in order to grasp and fully understand complex information.

This is where data-storytelling may help.

In simple terms is a balanced mix of storytelling techniques and data-driven information built through a meaningful design process.


Some of the technique we can borrow from storytelling are:

  • narrative structure around a clear message
  • use (text or visual) elements to anticipate or set the context
  • sequencing chunks of information to give time to understand the whole
  • simplify visual elements to reduce cognitive efforts
  • injecting human or personal components to trigger emotional reactions to strength and enhance the communication level


Since the information and insights come from some data, here a list of the common task we're asked to face:

  • data are facts, we need to interpret them
  • data is material, we need to explore and manipulate it in order to grasp
  • bias are everywhere, even our own bias
  • "interviewing the data" or understanding how the data has been collected is mandatory
  • deal with the statistically significant value
  • data are not perfect, errors or missing is the norm
  • context is everything, we need to get it
  • the sources need to be verifiable
  • every conclusion might be wrong, double test everything
  • correlation is not causation
  • scales are powerful and they can trick the perception
  • sloppiness is also a source of misleading


The process is the glue of the above parts. Nowadays, it needs to be an iterative process based on these main steps:

  1. explore/define
  2. prototype/test
  3. review/fix
  4. go to 1


Trying to define data-storytelling is not easy. It's a combination of many different techniques and strategies. One size to fit all recipe is not possible in this context.

Nevertheless, due to the complexity of the today information world, we need to push the limit in the communication field by means of new breeding techniques and artefacts.