Aristotle defined three modes of persuasion used to convince audiences: πππ‘π¨π¬, ππππ‘π¨π¬ and ππ¨π π¨π¬.
How is this related to Data Science? A great part of a Data Scientist job is to know how to tell a story, how to persuade or convince stakeholders that your results are reliable. Today I wanted to share with you how to storytell your results to the stakeholders of your company, one of the most important steps that you have to do to help your organization (focus on your business or ideas) understand the potential and usefulness of the data, by simplifying the complexity behind it.
As you may already know, Data Science is not only about generating valuable insights from vast amount of data. Two-thirds of the total amount of time that a Data Scientist spends a day working, is put on creating and communicating the results. This might seem easy, but it is the most important task a Data Scientist has to do, and should be executed in the most precise way.
When I talk about storytelling, I am not only talking about presenting clean and good looking results, which I consider that some people in the field tend to care too much about the outlook of the results and not about the content itself. When you want to create a story from the data, it is fundamental to understand what is the meaning behind the data and what is the data telling you directly.
The key behind storytelling relies on identifying relevant questions to be answered and communicate the answers to the stakeholders, or basically persuade them by reasoning or given strong arguments learned from data. Your main objective is to answer question such as what, where, who, what, why, and how.
Once you have a solid understanding of what are the opportunities for using that data within your organization, then a research-based approach should be preserved in order to sustain your βstoryβ. This is where the scientific background of the Data Scientist comes into play, and why is important to have it. A good story should follow the next steps:
β’ Make sure that the context is up to date, reasons behind why you are undertaking that analysis, always keeping in mind the questions that you want to answer depending on your organization and research on what has been done before (related work).
β’ Explain the reason behind the methodology used, if there is one ore more methods used in the analysis, explain the reason behind the choice of methods.
β’ Present the results obtained from the data in the most clean way.
β’ Discuss your results based on the prior research in the field. It is important to keep everything very transparent, if was not possible to generate good results, you need to communicate that and ways of solving the problem. If necessary, come back to the first point, and review.
β’ Finalize with a clear and concise conclusion, stakeholder would be in general mostly interested in this part, as for them is what would βgenerateβ value within the organization.
With all this in mind, as a Data Scientist you should be able to simplify the complexity of a problem, and make them interesting for your listeners (stakeholders). Your insights will eventually be the key for your organization in terms of decision making.
So far seems like an easy task, as we have steps to follow. Well, the difficult part comes when we have to start being creative and showing the great communication skills that we have acquired with previous experience. A strong narrative is necessary to explain, enlight, and engage your audience. And now you might be asking yourself, how can I improve these skills?
Donβt worry, everything is possible to improve with time, patience, and effort. I suggest reading and writing as much as possible (as I am trying right now π ), donβt make boring reports make them cool and interesting, and understand the needs of your audience so you will be better at persuading them.
