How do you tell stories with data?

Think of data storytelling as telling a tale, but with numbers. It's not just throwing charts at people. It's about weaving a clear, engaging story using data as your main characters. Just like any good story, it should grab attention, make things clear, and leave an impact. Whether you're sharing insights with your team, convincing investors, or teaching, data storytelling makes complex info easy and interesting. It's like turning numbers into a captivating novel!

In the ever-evolving realm of data science, where countless professionals can run similar analyses or build predictive models, what sets an individual apart is their ability to make data relatable and actionable for their audience. Being a great data storyteller means transcending beyond the numbers, to the heart of what those numbers signify. A well-constructed data story can simplify complex concepts, spotlight critical insights, and drive strategic decisions. As data continues to proliferate across industries, the demand for professionals who can not only analyze but also effectively communicate these findings will surge. Mastering data storytelling is not just an added skill; it’s a pivotal differentiator in the data science world.

One thing I have found useful is to look for patterns in data to find an anchor for your narrative. Contrary to popular belief, data storyboarding skills is not limited to only those who are tech savvy. It is a skill that can be built with consistent application of expertise and perspective and chiseling of edges to carve out a simple, yet comprehensive narrative to enable decision making.

Data storytelling is a potent tool in organizations. A well-crafted data narrative can shape a company's strategy, those who do it well understand the blend of art and science. Key elements to good storytelling include: ? Audience awareness - keep it focus and concise ? Burning platform - why should someone pay attention ? Impactful visuals - one visual can summarize an entire story in a split second ? Proactive guidance - Anticipate the needs of your audience In my career, I've leveraged these guidelines successfully in my data storytelling. Mastering data storytelling is not just about conveying information; it's about influencing decisions and driving meaningful change.

Data Storytelling is just another fancy term for "normal" storytelling. Please do not fall into this fallacy that you need to learn this. Learn the fundamental of how to tell a good story: Pay attention to any good movie: 1. start with a hint of why you should pay attention 2. give the context (why we should care) 3. show the trouble 4. show the journey 5. profit

Data in its raw form holds little value unless it's transformed into a compelling narrative that can be communicated effectively to end-users. The ability to analyze, integrate, and process data meaningfully is essential for its presentation. Data storytelling plays an important role in understanding the importance of data, its workflow, and the insights it offers to end-users. This skill is in high demand, as not everyone can adeptly convey the narrative, particularly when dealing with raw data.

Consider a health study showing 30% of participants are inactive. Alone, it's just a number. But add a story: imagine Jane, a mother who's inactive due to her sedentary job and feels too tired to play with her kids post-work. Suddenly, the data has a face and emotion. Data storytelling isn't just presenting numbers; it's about connecting those numbers to our shared human experience. Just like how novels capture our hearts, data stories drive real-world actions by making abstract figures deeply personal.

I think of data storytelling as the bridge between numbers and insights. You really have to be good at it as a data scientist or an analyst to be able to give actionable insights, otherwise your work and your results may mean nothing to a non technical audience.

After understanding the complex concepts about data, most of the data scientists believe that their colleagues from other departments, decision-makers can easily understand what data explains. Regrettably, I also attempted to explain how tests function, what specific values signify, and so forth. With more experience, I've come to realize that one of the most critical aspects of being a data scientist is the art of data storytelling. Currently, my focus is shifting towards emphasizing how to derive value and insights from the data results, rather than dwelling on the specific techniques employed and the underlying logic.

The purpose of using data to tell a story is not merely to present information, but also to assist the audience to gain insights and make informed decisions. Effective data storytelling combines data analysis and visualisation with a narrative that resonates strongly with the audience.

Imagine you're a city planner, aiming to promote bike usage. Your audience? City residents. Your data shows bike lane usage over a year. To find the story: identify trends and outliers. Perhaps in May, bike lane usage spikes. Coincidence? Not when you realize May hosts a popular bike marathon. By highlighting such patterns and events, you make the data relatable. Your story isn't just about bike lane stats, but community engagement and promoting a healthier lifestyle in the context of city events.

When I build data stories, I try to think about why people care about this, usually focusing on fellow engineers, data scientists, and managers. I typically write it as a slide deck, using statements as placeholders like, "Here we'll have a graph of the daily growth of new user sign-ups." Once you have those placeholders, creating the visuals and running the queries become the easy parts. Having someone review the proposed deck before you do the heavy lifting of running all those queries or analyses can save valuable time.

Define Purpose: Clearly state why you're analyzing the data. Gather Data: Collect relevant, clean data from reliable sources Explore Data: Use tools like charts and summary statistics to get an initial sense of patterns and outliers. Hypothesize: Formulate hypotheses about potential trends or insights within the data. Dig Deeper: Dive into specific data subsets, conduct statistical tests, and identify correlations. Visualize: Create graphs to make patterns more evident. Tell a Story: Narrate around your findings. Explain the context, implications, and why it matters. Iterate: Refine your story through feedback and further analysis. Engage: Present your data story in an engaging way to your audience, emphasizing the main points.

After doing a data analysis, it's often useful to go back to one or a few cases. For example, if the data is about people, are there a few cases of people that you can go into detail to convey the overall meaning and importance of your analysis?

Here are some tips for finding a data story: 1. Start with a question. What do you want to learn from your data? Once you have a question, you can start to look for data that can help you answer it. 2. Consider your audience. Who are you trying to reach with your data story? What do they already know about the topic? Tailor your story to your audience so that it is relevant and engaging. 3. Find the unexpected. What in your data is surprising, interesting, or newsworthy? These are the things that will make your story more compelling. 4. Use visuals. Visuals are a great way to communicate data in a way that is easy to understand and remember. 5. Tell a story. Tell a story that is engaging and informative.

Structuring a data story is crucial to effectively communicate insights and findings derived from data analysis. A structure that I’ve found helpful is the following: a. Introduction: Provide context for your analysis. Why was the analysis conducted? What questions does it aim to answer? etc. b. Data Presentation: Present the data visually using charts, graphs and other visualisations. c. Insights and Findings: Share the main insights and findings from your analysis. Use clear and concise language to explain what the data reveals. d. Supporting Evidence: Provide evidence and details to back up your insights. Include relevant examples. e. Conclusion: Summarise the key takeaways and reiterate the main message of your data story.

Best tip I every got on this was ,"Don't bury the lead". There is a reason you did the analysis and there's a specific decision you are trying to influence with your findings. Lead with that and then present your evidence as to how to got to that conclusion. This is going to be useful especially when your presenting to higher executives who may only have 10-15mins for your presentation.

Consider a public health campaign on reducing soda consumption. Hook: 'Did you know the average person drinks 40 gallons of soda yearly?' With visuals, the body demonstrates the rising trend in soda consumption linked with health problems. Insights show regions with higher soda sales have higher diabetes rates. Finally, the call to action: 'Join our 30-day no-soda challenge!' Here, you've taken data and woven it into a narrative journey from raising awareness to motivating change

One common structure for data story is: 1. Introduction: Introduce the topic of the story and the question for which we need answer 2. Rising action: Present the data and start to build towards the answer 3. Climax: Reveal the answer or the main insight from the data 4. Falling action: Explain the implications of the findings and what they mean for the audience 5. Conclusion: Summarize the main points and leave the audience with something to think about

I would suggest a five-step process for effective data storytelling that somewhat worked for me. 1. Know your audience and set up the context accordingly - what is to be expected? If your stakeholders are non-technical, no point talking about technical jargon. 2. Choose visualization elements that are intuitive not something fancy, so as to avoid the diversion of attention from story to chart explanation. 3. Don't pull all your insight cards all at once rather make it sequential as one leads to another. 4. Don't use too much text or make it too much colorful. It breaks the attention flow as people start reading rather than listening to you. 5. Finish it with a call to action like suggestions for improvements or further exploration.

Additionally, always remember Tufte's principle of maximizing the data-ink ratio. Eliminate non-essential ink to make the crucial data stand out. A clutter-free chart communicates more effectively.

Visualize your data story by choosing appropriate charts, simplifying designs, ensuring consistency, creating a visual hierarchy, providing context, using interactivity thoughtfully, arranging visuals logically, balancing complexity, considering your audience, and testing and iterating for optimal impact.

When it comes to visuals, less is best. The best charts are simple, clear, and effective. Refrain from using too many colours, or trying to put too much into a single chart. The chart should be easily understandable at first glance - nobody wants to stare at a chart for 15 minutes trying (and failing) to figure out what it is meant to convey.

Consider music. Music is a universal language. The arrangement of musical notation to create music can be understood and played by any musician, irrespective of the language that he speaks. In the same way, the right kind of data visualization can be understood by just about anybody who understands business or the use of data, irrespective of the language that he speaks

Before visualizing, thoroughly understand your data. Know its structure, variables, and the story you want to tell with it.Select a chart type that best represents your data. Common types include bar charts, line charts, scatter plots, pie charts, and heatmaps. The choice depends on the data's nature and the message you want to convey.

Our brain is naturally inclined towards good stories. Data stories are no different, if you can make them relatable then it is powerful! I will try to write one explaining Random Forest in simple words. Think of it as an election process where every citizen has their own considerations to vote and arrives at the final decision by evaluating them step by step. Our individual decision represents a decision tree that collectively creates the forest. Then what is random? different considerations! Think of a big bag of things we collectively care about to choose our leader and we all are choosing the ones we care most. That makes each of our decision as much uncorrelated as possible. As we aggregate all the decisions we get the chosen one.

Telling a data story is a crucial aspect of data presentation. It's not just about storytelling, but it's also about being authentic and transparent. Sharing the full story, even if it includes challenges or setbacks, builds credibility and trust.

One framework I have found useful is: 1. start with understanding the context ("who" is the audience, "what" do they need to know and "how" or what data are we going to use ? Craft a summary of thebidea to set context 2. Choose the right chart, focussing on reducing clutter. 3. Highlight part of the grapgh where you want your audience to focus. Example: highlights part of a line graph to show a change in trend after an event. 4..Make the design accessible with text and alignment to improve aesthetics. 5. Tell a story to tie everything together and end with a call to action. example: if a line graph is used to trace pricing of competitor products, build the graph to represent the state at different points in time.

To narrate your data story effectively, speaking your audience's language is essential, especially for those without a technical background. As the saying goes, "When in Rome, behave like the Romans." Even if you've created a brilliant solution, without the right presentation, it can fall flat. A good data story strikes a balance between industry knowledge and technical expertise. Your presentation should always reflect your unique style and personality, making it engaging and memorable.

Think of data storytelling as telling a story using numbers. Imagine you're diving into Flipkart, a big online store, to learn how customers behave during sales. Start by having a clear goal and getting your audience, like smart business people, marketers, or data experts, together. Get data about sales, customers, and what they like, and use tools to understand it. Make your story with care, starting with an interesting beginning, sharing important ideas, using nice pictures, and ending strongly. Show your story with pictures like charts and graphs, and explain why knowing what customers like is important for making decisions and marketing. Practice and get feedback to get better at storytelling, so everyone can understand data.

How to tell a story through data? Nothing different from the stories our granny's used to tell: - Know your audience - What problem are you solving for them? - Articulate the outcomes upfront. - Prove with your data and insights. Ex - Imagine you have a product in your hands which can save you an hour daily in auditing your team's work. Imagine you don't have to answer customers' nasty calls on some errors. Here's presenting a ML product that will audit your team's work automatically before it goes out to customer. Then the data and logics...

Data storytelling is an art of showing facts and in a way that has power to make your audience believe in what you are saying. Oftentimes it’s considered as a skill required for a few like people working as data scientists, analysts and engineers. But in fact it’s necessary for everyone in many situations in life.

Elevating stories is pivotal in our journey toward a data-driven AI culture! As a CDAO, I've witnessed the magic that unfolds when we shift from narrating data to orchestrating a data story. It’s not just about 'What Happened' - that’s a rearview mirror. It's about a comprehensive narrative that encapsulates: What Happened (Descriptive), Why it Happened (Diagnostic), What’s Going to Happen Next (Predictive), and What Should We Do About It (Prescriptive). True Data Storytelling takes us into the future where data and insights drive decisions. Where raw data transforms into actionable insights, creating a culture that is not just data-informed, but data-inspired! Don't just tell stories WITH data but let your DATA talk and TELL THE STORY.

Create a hook when you tell stories with data: When telling stories with data, it's crucial to create a hook that captures your audience's attention. You can achieve this by starting with a surprising statistic, posing a thought-provoking question, sharing personal anecdotes, or setting up a mystery. Alternatively, you can highlight contradictions, invoke curiosity, use compelling visuals, or relate your data to current events. Shocking quotes or "before and after" narratives can also be effective hooks. The goal is to engage your audience from the outset, making them eager to explore the data-driven story you're about to unfold.

Is the data that you rely on suitable for the targeted audience? Take a top-down approach and decide what business case you are trying to address. What type of data is at your disposal (data at rest, real-time data, uncertain data), and how much do you trust this data? What story do you want to compile based on this data (insights on what happened, why it happened, what will happen, or how you can make it happen)?