What's the best way to analyze data for storytelling?

Knowing your audience is absolutely critical for success in your analysis and storytelling. For me, there are four critical dimensions that will help you to better understand your audience and deliver value. First, you need to understand their problems or challenges. Second, you need to understand what outcomes they’re trying to achieve. Third, you must understand where they’re spending their time and resources (i.e., what actions, activities, or priorities are they focused on?). Fourth, you need to understand what measures or metrics are important to them. The better you understand all four of these dimensions, the more likely you’ll identify meaningful insights and tell data stories that your audience will care about.

Data storytelling is more dangerous than fiction. Why? People will openly question your story. They will doubt your methods. They will doubt your numbers. They will doubt your interpretation. Accordingly, make sure you know the 'story behind the story' - not just the top line numbers.

In writing, it's best to address a specific person or group familiar to you. Addressing to "everyone" is a terrible mistake made by many beginning writers

I believe that the selection of data for analysis must be driven by the target audience. Proximity and relevance are key to effective audience engagement hence the set of data picked for storytelling must fit the narrative that is interesting to your target consumers. Audience is king and anything done that is not in its interest is a straight flop. It is imperative that we first know our product consumers before packaging stories and visuals for them.

It's about knowing your audience. Unless you are writing for an audience that is likely to be numerate, data storytelling has to be simple and very visual. Analysis still has to be a deep dive, but all the work gone into that will be the part of the iceberg beneath the surface.

While you can have a hypothesis or a hunch when you begin your analysis, I’d rather listen to what the data has to say before defining a goal for my data story. What if the data disproves what I defined as the goal of my story? After you’ve explored the data, your intended narrative might shift in an entirely different direction. You never know what you will discover. I’d recommend being open-minded in your analysis phase before setting any goals for your stories.

One of the fundamentals is to keeping it simple. With data, learn to find a common problem or narrative that can be weaved to connect with your audience. While most data scientists and analysts love their own data to 'impress' upon the publics, it helps by removing the noise and focusing people's attention on the key insights. It requires building a narrative and revealing data in the proper order and sequence, as well as visualization.

I think there is an element of having a goal in mind, for the purpose of starting, but the analyses and findings should lead the story.

It's important to actually first know yourself that why the story you want to tell is impact, what change or what experience you want the reader/receiver to have. This could be just for information or driving an action, in case of an/ multiple actions then what you want to drive? In case of driving action are there enough cues you are providing?

In the realm of storytelling, your goal is your guiding star. It's the message you want to etch into your audience's memory. To embark on the data storytelling journey, you must first define this goal meticulously. It's akin to deciding the moral of a fable or the climax of a gripping novel. Your goal shapes the entire narrative, from data analysis to storytelling techniques. It must be crystal clear, specific, and measurable, like a treasure map directing you to your storytelling destination. Your goal isn't just your destination; it's also the compass that ensures your data story hits the mark, leaving your audience not just informed, but moved to action or thought.

When I do analysis, I try to find insights, not stories. My favorite definition of an insight comes from author Gary Klein who said it is “an unexpected shift in our understanding about something.” We believed something and then new information turns our prior understanding on its side. The best data stories don’t contain a bunch of facts and figures (information)—they feature at least one actionable insight that will rock an audience’s world. If you look for stories in the data, you’ll find them. The danger is they may be narratives that don’t matter to your audience, or you may inadvertently contort the facts into a desired narrative.

Every dataset can yield more than one inference (I hesitate to call them 'insights' at this point). Without a direction, data analysis can be tiresome and seemingly futile. This approach, known as Exploratory Data Analysis (EDA), at best, reveals patterns, anomalies, and potential relationships between variables. In contrast, Hypothesis-Driven Analysis begins by defining the exercise's direction, considering the specific needs. This method allows one to identify the appropriate data, analyze it, derive insights, create suitable visualizations, and ultimately craft a compelling narrative for the intended audience. Even the outcomes of EDA can be used to formulate hypotheses that can lead to a compelling narrative.

Data can be quite extensive and trying to figure it out can get overwhelming. The best approach is to note down which data best relates with your goal. You can then list all the elements that affect this data. To ensure there is no bias, you can test a few hypotheses and then move ahead with your data story with confidence that you are going in the correct direction.

Data in today's world is easy to find but hard to interpret as the absolute truth. That is why it is important to gather data from 1. Trustworthy sources 2. Multiple sources Use your discretion then to analyse and find the pattern because more often than not the true facts come out when data from multiple sources are analysed together even if in silos they could be contradictory to each other! I have often also relied on non-digital sources like research papers and books to validate some facts. Modern storytellers often forget that traditional methods of data mining like making notes from written documents works well even now and are a very reliable source of information.

Explore the data and let it lead the story with its results, using this techniques is it helping you understand something better, diagnose a problem, find a solution or change a mindset to be able to take action?

This is the area where I see the most confusion and the most bad advice. Having a good narrative structure to follow helps you turn your analysis findings into a coherent, powerful story. The suggestion that stories have ‘a beginning, middle, and end’ is almost worthless. Other frameworks are more geared toward executive summaries than stories. My preferred framework is based on Freytag’s Pyramid. You establish the setting/context before introducing a hook (a key observation from your analysis) and then unpack that observation until you reveal your big takeaway (aha moment/main insight). You then conclude the story with a recommendation and next steps (call to action).

This is the part people struggle with the most and can be different for the different purposes or audiences. Finding your flow is also really important in your authenticity of storyteller.

Crafting narrative is exciting, as it is creating the message one wants to convey. There is a possibility that one's personal biases or cultural conditioning may play a role in this. It may or may not be appropriate, however, it would be a good a idea to have cross cultural understanding or at least try to build one. Having worked in large teams across different states, I used to make sure that I keep myself updated about cultural and cognitive beliefs of the communities about whom I wanted to talk as well as about the recipients of my story.

Once insights are derived from data analysis, constructing a compelling narrative becomes crucial. A narrative, following frameworks like the hero's journey or the three-act structure, provides a logical sequence connecting data points with a defined beginning, middle, and end. Ensure your narrative includes key elements like a hook, context, challenge, solution, and a call to action for maximum impact and engagement.

A few models that I use based on what I want to showcase through data. 1.Inverted Pyramid: Used in journalism, where the most important hook is presented first, followed by details in descending order of importance. 2. SPIN Selling: A framework used in sales to uncover a prospect's Situation, Problem, Implication, and Need-payoff. 3. Feature-Benefit Framework: Presenting product features followed by corresponding benefits to the customer. 4.Monroe's Motivated Sequence: A persuasive speech framework that includes Attention, Need, Satisfaction, Visualization, and Action steps. 5. Chunking: This is a more modern type that works for social media short stories. Use two data points and create hypothesis or vision around it.

The best format which I’ve seen working well is — PowerPoint presentation which includes SIMPLE graphs. Reports are generally dull in nature. With PPT, you can have main point highlighted and explain it in detail verbally. The graphs which is used for data visualisation should be simple and understandable, even if you don’t explain it! Complex graphs may make you look like you’re a pro, an expert who understands complex stuff. If you want your listeners to leave with takeaway, keep the graph simple, and you’ll be known for good storytelling ??

One of the common misconceptions lies in confusing data storytelling with data visualisation. data storytelling is all encompassing and includes data visualisation. The narrative precedes the visualisation. There are so many ways to bring alive the story through appropriate visuals , infographics etc. the equivalent argument is in making an attractive slide deck losing track of the core message ( story) that needs to be communicated.

Choose the format in which you can best represent the message you want to convey. Think from the perspective of the audience to which you want to address.

Simplistic hand-drawn doodles can represent the core message and value of a data set. For instance, a few ways in which I showcase are: 1. Descending arc 2. Ascending arc 3. Nature-inspired Doodles - Sizes 4. Quadrants to represent movement.

The three keys to Storytelling I have experienced: 1. Simple- to resonate and understand 2.Impactful - with qualitative and quantitative needle movers. Data visualisation helps. 3. Memorable-it is shared and quoted as an anecdote by others in future conversations.

Trying to figure out WHY a story doesn’t FEEL right is hard. Check to see it has these narrative elements: 1. HERE’s WHAT 2. SO WHAT 3. NOW WHAT Next, do an audience check. What should the audience be… 1. THINKING 2. FEELING 3. DOING …differently as a result of your communication? If you can’t identify these elements clearly- keep workshopping until you can. In my communication planning, I usually start with DOING after I’ve performed my data analysis and work backwards from there as DOING and NOW WHAT usually are very similar. Analysts: Step into confidence- if you did the analysis- you have the PROFESSIONAL AUTHORITY, ability, and responsibility to provide clear actions others more senior than you COULD take. Share that NOW WHAT!

Make friends with those whose experience sits in their field, journalists, content designers, copy writers, they can offer real test to the story and give you advice to hone your skills.

An example of how to test and refine for video storytelling is through social media data analysis. Say you upload a 3 minute video to your brand YouTube page and directly to LinkedIn. Which platform has higher CTR? Engagement rate? While platforms may differ in how they measure a "view", looking at completion rate is a good measure of storytelling success. Did this video beat your average video completion rate for videos of the same length? Did it beat the industry average? If not, look at where they fell off and work backwards to determine what is missing in your story. If viewers drop off at the 30 second mark when the video starts explaining a products features rather than continuing the story, now you know to cut that section.

I learnt how to conduct ethnographic research during my master's degree, and it taught me a lot about presenting qualitative data. It is focused on a method called coding, which helps one detect patterns or sequences in the data. These, in turn, can help a researcher identify themes and build theories. I use this method for sorting qualitative data in my content writing work as well, for example when interviewing people or during data-collection through a survey. I code the answers received based on the words people use in their responses, and see how that correlates with the theories that exist, and phenomena that present themselves. This method strengthens your arguments, and limits bias to a great degree, so I highly encourage its usage.

As much as I love data, I think the really useful stories comes outside the data. Here's what I would suggest: 1. Understand the business domain by partnering with the subject matter expert. 2. Listen to the customer interviews to understand their problem areas 3. Look at the customer feedback portal 4. (If you can), Interview some of the customers you see in your data When you add all of these and combine with what you are seeing in the data, makes for a compelling case. This will help you become a better Change Agent!

The creation of a single world comes from a huge number of fragments and chaos. Data reduces the chaos and it numbers the fragments. If that's your thing, go for the data. A paraphrased adage to encourage you: Write for everyone and be read by no-one. Write for a single person and chances are, you might be read by everyone. Something to dissuade you: Everything that defies a genre was raw, gritty and original. You can simply express yourself with or without data. Your choice. :)

One final thought. If you can write this story in a way that captivate both people who understand numbers, and those who don’t, you will have a much more compelling story. Allow the readers to understand the data, but not feel like it is the centerpoint of the story. Allow your storytelling to push your readers in the direction of the numbers.

I collect beta and alpha reader feedbacks (books). One thing I focus on the most is where did people pause*. The pause comes due to: confusion, boredom or too much excitement (which leads to skimming). Find the commonly mentioned pauses, fix those places. It works incredibly well. *A pause is different from negative space - breathing room for readers to comprehend what happened.