Being the Casey Neistat of Data
On good data storytelling and how to possibly get there ...
Preface
This is a follow-up post from the previous one on “Questions First, Tools Next”.
We recently covered the topic of Data Storytelling in one of my classes at school. I’m not unfamiliar with the term. The funny thing is I mentioned wanting to be a Data Storyteller in one of my college interviews (NTU, if you’re curious). Who knew I’d come across the term again year in!
The same night I came across Casey Neistat’s Instagram Story. He was in some random desert driving with a friend. No big deal there, he’s crazy and spontaneous like that. But, it got me thinking back to the lesson on Data Storytelling.
A nice takeaway from the lesson is to be the “Casey Neistat of Data”.
Interested in my free form writing? Consider signing up to my newsletter:
Casey Neistat
Most of you reading this probably don’t need an introduction. To those who do, he’s this amazing vlogger on YouTube based in New York City. He’s an expert at telling stories that matter, stories that inspire, and stories that have calls to action. His work has gone viral multiple times owing to his ability to create connections with his viewers. When he speaks, it’s as if he’s literally in front of you. I’ve been watching his videos since the days of Dollar Pizza, Beam, and Merlin the Mailman – basically forever. Sometimes, his vlogs help me catch up on the regular dose of NYC.
If not for the average person, I’m sure the tech bros reading this would fangirl over Casey any day. Well, I’m in that bucket too.
Casey is known for his slick montages, drone footage, monologuing, cinematography, and sharing of life lessons. Let’s think of these as his media, which by the way, he uses really well. Though, ultimately, his viral videos all stem from the same messy source – his camera.
Sure, you can argue he has some great equipment now but mind you, he started off with a Camcorder. Regardless, he has this “other thing” that makes him a good storyteller. It’s this little secret sauce that adds spice to the story.
YOU. NEED. EMOTION.
From what I can tell, he’s able to use media to draw out emotion from the viewer. You need the users to feel something and this can be anything – happiness, extreme sorrow, frustration, literal THRILL. You name it. He knows how to tag each scene (or set of scenes) with a particular emotion. To achieve this thing called “evoking emotion”, you’ll need 4 things according to my analysis:
Purpose – why are you showing a specific scene/montage? What’s it supposed to evoke?
Order – what’s the specific order of scenes? What kind of story are you trying to tell?
Element Format – what are the different components used (drones, montage, timelapse, monologue)? Is there variety in the types of scenes being displayed?
Significance – why is the scene important to You, the creator, and your viewer? Is there a call to action or a positive takeaway?
Once you have POES, you can easily stitch together some great work (of course, some level of skill with the tools is necessary; more on this later).
At this point, I’d like to put up a disclaimer that this is not a get-viral-quick ploy. Virality is a whole new issue I don’t want to get into. Besides, if I could do it, I wouldn’t be sharing it anyway.
Also, you have to try and fail at this multiple times before you succeed . You cannot be a one-hit-wonder on YouTube that easily. Casey Neistat didn’t become The Casey Neistat on the first try.
Casey has figured out how to use POES to brand himself and his videos. And the rest is history, as they say.
Back to Data Storytelling
Now think of Casey’s camera memory as the internet and the raw footage inside as the data you’re trying to explore. It’s messy, unorganised, and has no meaning on its own. Now, think of Premiere Pro (his video editor) as the tool that helps shape and mold that raw, yucky camera footage into the masterpieces on his YouTube channel.
Long story short, being a data storyteller is no different!!!
Storytelling is all about playing with the viewer’s emotions (in a good way). You should know when to make them rise and when to bring them back down again. Do this flipping up and down frequently and you have a thriller or action scene. It keeps them at the edge of their seats. Do this infrequently and get a feel-good sombre scene. You simply need to think about POES at each stage of a Data Science/Analytics project (like this guy’s TED talk here). Every single thing you write and figure you display must have an intended purpose, it should appear in the right place, it should be displayed in the right format (graph type, axes, labels, colors, font sizes, etc.), and should be important to the reader in some way (why should they care?).
This, to me, is the core essence of data storytelling the Casey Neistat way. This is how you create compelling stories that are personal – and it applies to data in the exact same way it does to YouTube.
Asking “So What?” Often
This is great advice I was given by my seniors and mentors over the years when writing essays. For every single sentence, ask the question “so what?” at the end. This helps you understand whether you are explicit about what you want to convey. Often, writers miss out on details because they assume readers know it. While it’s nice to assume an educated and aware audience, explicitly clarifying things is great depending on the context; it helps you remain concise.
The same philosophy of “so what?” applies to data storytelling too!
For every single sentence you write, or figure you add to your quantitative reasoning/data science/data analytics story, ask “so what?”. This question is your North Star and guides you to write what matters. You should stop when it gets too difficult to answer the question or you might risk becoming too longwinded. If you are unable to answer it, you probably are unsure of your own analysis OR you used the wrong means to display your content (i.e., one or more of the POES components may not be the best).
Data Storytelling is Iterative
I want to circle back to an article written by my close friend, Sarv titled “The Tyranny of the Faceless Other”. It’s a great read and contains many lessons he and I learned the hard way, on our own different paths.
In the essay, he advises not to chase perfection from the get-go. It’s easy to suddenly stop without inspiration because your work isn’t up to par with whatever standards you or someone else set. More often than not, it’s you who sets unrealistic standards for your work, causing this sudden loss in productivity and motivation to continue.
Data Science and Quantitative Reasoning can never be perfect from the start. They are iterative processes.
Most times, when writing an essay or building a project, we try to immediately finetune and optimise every minor detail till it looks “pErFeCt” in the first go. I’ve been advised to start off making the crapiest thing ever. Seriously, once you understand your data (columns, labels, values, etc.), create some basic figures and write down basic observations from them. Next, pick a specific figure or block of text (i.e., your analysis or “story”) and add more details to it. For figures, try out new plotting styles or graph types. From here, improve each and every component.
As the storyteller, you know best what you want to do.
The thing about creating QR projects or essays is that they are autoregressive i.e. they are built sequentially or iteratively over time. The opposite of this is non-autoregressive where you expect to build your project perfectly at a single shot without any reattempts, and this is plain impossible.
All About Playing Around
Do you really think Casey Neistat gets the perfect video on the first try when he sits down to edit? NOPE!!! That’s what the internet makes you believe when you see his work on YouTube. It shows how much grind work he puts in behind the scenes to iteratively polish his videos. I can assure you it’s incremental progress, a lot of reviewing, and a lot of shifting around.
It’s about playing around with what you have. Playing around is fun, and fun is what you’re trying to achieve with storytelling.
Before you engage with data storytelling, play around with your data. Try to find interesting patterns that you’d want to convey to your audience. Definitely, the more you do it, the easier it gets. You start forming personal philosophies for your craft as did Casey Neistat before his prime.
In a Nutshell
If I had to wrap it up with a bowtie, data storytelling is about playing around with data. It’s about POES and relevance. It’s about telling a single other person why you find XYZ exciting in the most enthusiastic way possible – and you do it all through text, charts, images, and analysis. That single person becomes two, two becomes four, and so on.
I hope this hasn’t come off as a “demystifying YouTube” article or “get viral quick” scheme. You can think of this as an accumulation of lessons from across the past few weeks of taking a QR class at school. It’s just one more possibly-interesting story I wish to share.
It’s been a wonderful 12 weeks with this class. I thank Prof. Charles Burke for his wisdom, his “calling out” behaviour that I’ve found amusing and useful, his fashionable sense of humour, and data storytelling skills he has carefully taught us.
With that, I’ll see you in the next one!
My Call to Action … Of Sorts
Like what you just read? Consider signing up for my newsletter! I talk about Machine Learning, Artificial Intelligence, Data Science, and Technology in general.
Love your article! Experimenting and playing around with data is crucial to train the eye in identifying interesting patterns as well as narrating a story behind these datasets to your audience. Absolutely agree with you that the more we do it, the easier it gets!