Storytelling with Data: Results Explanation and Storytelling

The final stage of the OSEMN framework — interpreting (interpreting) — is where all previous work has taken place. If you can't deliver even the most excellent analysis, it doesn't turn into a decision. In this article, we learn how to explain the results and turn the dataset into an impressive story.

This post what is data analytics a continuation of its guide.

Interpretation stage: turning numbers into stories

At this stage, the goal is to turn hard numbers into a story that everyone in the team — not just data experts - can understand and act on. This is a critical step: having an opinion is one thing, delivering it effectively is another. Choose the correct medium first — a slide presentation, interactive notebook, or detailed report. Each has its strengths and weaknesses.

Choosing the right medium

An important part of effective data storytelling is choosing the right medium to communicate their findings. There are several options, and each has its own strengths and weaknesses:

• Slide presentation — universal in all industries; structured, visual and engaging. Ideal for a large audience.
• Interactive notebook — for a technical audience; displays code and output in one place.
• Detailed reporting — for guidance or formal review; when depth is required.

Whichever medium you choose, the principles of storytelling remain the same. In this article, we focus on a slide presentation because it's the most common format-but lessons also apply to other mediums. Storytelling is an essential skill in data analytics to adapt the approach to each medium.

Effective slide presentation structure

A slideshow presentation is a structured and visual way used across all industries. A good introduction should follow these steps:

1. Remind me of the original issue — what were we trying to fix and why were they important?
2. Highly display the method you used (OSEMN steps) — without going into technical details.
3. Present the dataset with a visual tour — show key data points with graphs, charts and tables.
4. Explain visuals — interpret what examples mean, in the context of a problem.
5. Submit recommendations — what steps should be taken based on the findings.

Nümunə: e-ticarət mağazasında son üç ayda sayt trafikinin ciddi düşməsini araşdırırsan. Təqdimatı problemin xatırladılması ilə başlayırsan, sonra OSEMN prosesini bölüşürsən (sayt analitikası datasını topladın, təmizlədin, araşdırdın, model qurdun). Sonra bir qrafik göstərirsən: trafikin azalması ilə səhifə yüklənmə vaxtının artması arasında aydın tərs korrelyasiya. Nəticə: səhifə yavaşladıqca istifadəçilər çıxıb gedib. Tövsiyə: saytın performansını optimallaşdırıb yüklənmə vaxtını azaltmaq. Beləcə data tam dövrə vurub konkret hərəkətə çevrilir.

Answering a business question drastically

The comment stage is to go back to where you actually started. The first four phases of OSEMN — gather, clean, explore, model — each deepen their understanding of the business challenge you're trying to solve. The goal now is to take the outcome of the model and clearly answer the original business question. The model can predict how many responses an email campaign will get, for example, or find new leads-but these numbers only gain value when interpreted correctly.

A good comment holds three things together: explain (explain — what the outcome means), illuminate (enlighten — why it matters), and engage (engage — move the audience). It's not just about showing a number; you have to communicate the meaning behind that number and what to do with it.

4 parts of an impactful story

From novels to films, there are four parts to every impressive story: setup (introduction), buildup (development), climax (culmination), and conclusion (conclusion). The data story must also carry these elements.

Setup — input and hook

Like every good story, start with a «hook» (hook) that attracts an audience — a question that often arouses curiosity. Example: Inu and Neku sales have seen a sudden drop in the past few months. Clear hook: why? What could cause this fall?

Buildup — development

Here's where the story unfolds: you explain the steps and findings you took to explore the hook. In this example, you see that sales are coming from internet, wholesale and retail channels; the fall is driven by internet sales. If you look at the site dataset: internet sales are down, but visitors are not. As you go deeper, you see that the cart abandonment ratio is high. This is the main point.

Climax — culmination

The moment you uncover the main reason for the hook — the instance where the audience «bulb burns down». Example: 92%of customers leave the cart during the purchase process, although this has not been the case in previous years. The reason is obvious: the products they want aren't in stock, so they can't buy them.

Conclusion — results and recommendation

You are introducing an action that will resolve the issue. Example: if we increase stock of required products, the drop in sales goes backwards. You don't always need a neat story, but the most effective discoveries often come in the form of a story-especially when data shows unexpected, complex, or expensive results.

Understanding model results

Nəticələri başqalarına çatdırmadan əvvəl özün onları düzgün anlamalısan. Model bir proqnoz və ya nümunə verir, amma bu, hələ cavab deyil — sən onu biznes kontekstində şərh etməlisən. Məsələn, model «səhifə yüklənmə vaxtı ilə trafik arasında güclü mənfi əlaqə var» deyə bilər. Bunu anlamaq sənin işindir: yəni səhifə yavaşladıqca insanlar çıxıb gedir. Model həm də qeyri-müəyyənlik göstərir (məsələn, 95% etimad) — bunu da nəzərə almalısan ki, nəticəyə həddindən artıq güvənməyəsən.

A good comment doesn't confuse causation with correlation. Two variables can change together, but that doesn't mean one causes the other. The data analyst needs to understand this difference and be careful when presenting results — otherwise they may make the wrong recommendation.

Full example: Inu and Neku sales fall

Let's combine the four parts of Storytelling into one example. Suppose that sales of the Inu and Neku store have suddenly fallen in the past few months. Setup: hook is a question «why does it fall?». Buildup: you investigate and see that sales are coming from three channels — internet, wholesale and retail; the fallout is from internet sales. If you look at the site dataset: sales are down, but visitors are not. As you go deeper, you discover the high abandonment rate of the basket.

Climax: 92%of customers abandon the cart in the buying process-it hasn't been in years. The reason is obvious: the products they want aren't in stock. Conclusion: if we increase the stock of required products, the sale returns. Here's an example of how raw numbers become a clear, action-packed story. Audiences now know not only the fact that “sales are up”, but also the cause and solution.

Notice that this story reflects the OSEMN era: data collected (site analytics), cleaned up, researched (breakdown by channels), model built (connection between cart swatch and sales), and finally commented and turned into a story. Storytelling is the final step that keeps the whole process together.

The role of visualization

The role of visuals is not just to show the dataset — it's to tell the dataset itself. Highlights good graphical trends, anomalies, patterns, and correlations. But after you submit the visa, you have to explain it: you are in the role of an interpreter, you decode the visa and make the dataset into a story. The explanations should be simple and cohesive so that the person who is not a data analyst can also understand.

Frequently asked questions (FAQs)

Why is data storytelling important?

Because it's not enough to have an opinion — if you can't deliver it effectively, decision makers can't act. Storytelling turns data into real decisions.

Do we need a story for every data analysis?

Not always, but when the outcome is unexpected, complicated, expensive, or surprising, storytelling is key.

Which introductory medium to choose?

For instance: slideshow for a large audience, notebook for a technical team, detailed guidance report.

Is the presentation to a technical audience different?

Yes. For a general audience, you can simplify details; for a technical audience, you can open more details of methods and models.

What are the parts of an impactful story?

The four parts: setup (introduction with hook), buildup (investigation and findings), climax (uncovering the underlying cause), and conclusion (recommendation).

Are correlation and causality the same thing?

No. Two variables can change together, but that doesn't mean one causes the other. The analyst should not confuse this difference.

What is the role of visualization?

The dataset is not just showing — it is telling the dataset itself, highlighting trends, anomalies and correlations.

What does explain, enlighten, engage mean?

Good commentary combines three things: explaining (what the outcome means), illuminating (why it matters), and engaging (acting on the audience).

What steps should a slideshow presentation follow?

Remind them of the problem, show the method (OSEMN), visualize the dataset, explain the findings, and end with a recommendation.

When to use data storytelling?

When the website is difficult to understand or the impact on the business is large. When data shows something unexpected, complex, expensive, or surprising, storytelling is key.

Engaging audiences (Engage)

A good data story doesn't just tell-it moves audiences. Here are a few principles. First, tell the story in the language of the audience-the person who is not a data analyst should also understand. Second, keep the visuals simple and clear; it drowns out an overly complex graphic message. Third, each slide has one key message — people can't grab too much information at once. Fourth, end with a specific recommendation so your audience knows exactly what to do.

Involvement is also an emotional connection. The numbers themselves may be dry, but they're memorable when presented within a story. “Sales fell 15%” is one thing; “92 out of 100 customers leave their cart because the product they want is not in stock” creates a whole other effect. The second encourages the audience to act immediately.

Result: turning dataset into action

Commenting and storytelling is the last but perhaps most critical step in the OSEMN era. This is the bridge between raw data and real decisions. Even the most in-depth analysis remains on paper if it cannot be clearly communicated to the people making the decisions. That's why it's important for a data analyst to have as much technical skill as the ability to communicate and storytelling.

Remember: a neat story isn't necessary every time, but the most effective discoveries — when data shows something unexpected, unpleasant, complicated, or expensive — usually come in the form of a story. In such cases, storytelling is not an option, but a necessity. It becomes invaluable for an analytical team that is able to translate data into the right story.

Practical advice: the next time you submit an analysis, ask yourself first — “what is the story that this dataset tells?» Find the Hook, set the Find as the main point, and end with a specific recommendation. If you practice multiple times, it will become a natural habit to turn numbers into stories and make presentations more impactful.

Next step: learn the new technology that accelerates this whole process — generative AI.

👉 Next Post GenAI data analytics