Data Storytelling 3: tone, questions to ask, storyboards
Back from Mexico, and happy to see a reader was able to apply the advice here and get the ball rolling on his career.
This post in a continuation of the last one, you can read it *here*
1 - Tone
1.1 Tone
What tone do you want to set? Depending on what you are communicating you’ll want to have a very different tone in your voice.
Are you celebrating a predictive model’s success?
Trying to sell a ML model to the suits?
Is it a lighthearted or serious conversation?
The tone of your presentation/voice will basically dictate what your presentation (data storytelling)’s structure will be like.
Here’s a useful video on how different tones lead to different conclusions:
1.2 How
Cool, so at this point you’ve figured out your audience, you’ve figured out, you’ve figured out what they need to know, you’ve figured out the tone you need to set for your presentation. Now, you’ll need to know the how, aka:
What data do you have available from your dataset that will help make your point?
Now, an amateur move (news outlets do this all the time btw) is to lie by omission, aka when you are presenting, you just ignore all of the data that doesn’t support your point. This doesn’t work in the real world, as all it takes is just 1 person who is knowledgeable, and the moment they poke holes in your presentation, the rest of the suits immediately question your credibility.
No credibility = No persuasion. Instead, it’s your job to bring up the data that doesn’t support your point, and explain why you believe that your thesis holds even when there’s some data saying you are incorrect (could be outliers, typos, etc…)
2 - Questions to ask
In the real world, a lot of time when you are asked to give a presentation, it is usually at the request of someone: C-suite, stakeholder, or a client. This means when you are trying to figure out how to give a good presentation, the only information you have is what your boss told you. Well, what happens if your boss didn’t get all of the context?
In this scenario, the best thing you’ll want to do is directly reach out to the original requester to fully understand the situation (the real question they want answered). Usually, what you’ll find is that there is additional context in the head of this requester that they assume you already knew, or they forgot to mention.
*Important Career Note*: Usually most bosses are fine with you directly reaching to the requester to get the job done, hell sometimes they’ll even be proud of you taking an initiative. But, sometimes you’ll get talentless middle
Keep reading with a 7-day free trial
Subscribe to Data Science & Machine Learning 101 to keep reading this post and get 7 days of free access to the full post archives.