Data Storytelling with Auto Insights

[MeganDibble]

Introduction to the 7 Data Stories

In 2019, the New York Times published materials from their data training program. In their tip sheets, they outlined the seven types of data stories. These categorizations helped them classify data for their articles and then structure the piece accordingly.

Alteryx Auto Insights (AAI) excels in telling all these data stories, so let’s look at each one. The examples in this article use a dataset where the measure of interest is Revenue (over time).

1. Change Over Time

One of the most common questions about a data point is, “how does this compare to the last period?” Periods could be any measure of time from days to years, depending on the business reporting needs.

What’s nice about Auto Insights is that it tells the “story” in graphs and plain language. In the image above, the sentence about the total revenue (near the top) is auto-generated. Forget your monthly PowerPoint summary—Auto Insights can report the whole message to your stakeholders!

2. Drill Down

In Auto Insights, you can drill down to explain the key changes that occurred over time when comparing segments. This allows you to answer the question, “how do each of the company sizes relate to changes in revenue?” We can see very quickly in the drill down view that Enterprise companies contributed the most to the overall increase in revenue.

This kind of data story is beneficial for root cause analysis. In my experience as a data analyst, this is always the most critical subsequent analysis for business users—understanding the “why.” And with the Auto Insights product, you can see the “why” in a few clicks, rather than having to create a second analysis or even launch a follow-up project.

3. Zoom Out

If you are presenting to executives, they might not be interested in the details. (Don’t take it personally.)

You can show them the big picture by zooming out to the Measures panel at the top of the “Discover” tab in AAI. The big picture story with this example dataset is a positive one—total revenue, clicks, and impressions are up from the last period. The color-coded arrow icons add a visual element to your story compared to simply showing big numbers on a screen.

4. Contrast

Contrasting a data point with the metric's minimum, maximum, and average values tells another interesting data story. In the search tab of AAI, you can create a visual like the one above by selecting your metric and a comparison segment. It will default to show you the chart that makes the most sense for the data, but you can also change it with a drop-down selection.

The bubble chart effectively tells the compare/contrast story—you can immediately see the percent change in revenue by industry and that the industry with the largest revenue is Entertainment.

5. Intersections

In the example above, there is an intersection in the line chart between July and August 2021. The segment with the most revenue goes from “Small Medium Business” to “Enterprise.” Identifying shifts like this can be critical for conversations surrounding business strategy and the efficacy of company initiatives.

This chart was also created in the Search tab of AAI by selecting a measure and segment—the only difference was that I chose a line chart to see trends over time instead of the snapshot in time that the bubble chart showed. With one drop-down selection, I was able to tell an additional story!

6. Factors

Pie charts are the classic example of a “factors” story. They show the part-to-whole relationship of a segment of data. Pie charts, however, are notoriously unhelpful—it is hard to compare slices of a circle and gauge relative size. Enter: the stacked bar chart!

The chart above shows the portion of total revenue made up of accounts greater than nine years of age and how that has changed over time. There is also an option to see the chart as 100% stacked (i.e., just comparing proportions and not total revenue).

7. Outliers

Malcolm Gladwell? No, not that outliers story.

The outliers section of Auto Insights, found at the bottom of the Discover page, tells the story of data points far outside your dataset's average range. Once again, we get a lovely one-sentence summary of what is notable about this chart: vocational and investment funds increased well above the average industry L2.

You can also click on “What caused this?” to zoom into a specific data point, adding another layer to your story.

Summary

Here’s how I would boil down each data story into one question:

1. Change over time: “How does this metric compare to the last period?”
2. Drill down: “What are the relevant segments affecting this metric?”
3. Zoom out: “What is the big picture story for this metric?”
4. Contrast: “How does this data point compare to the distribution of the metric?”
5. Intersection: “How does the change over time of this segmented metric compare to similar ones?”
6. Factors: “What percentage of the total metric does this segment make up?”
7. Outliers: “Are there any data points that are significantly different from the rest of this data?”

Data storytelling is an essential skill in the information age. Hopefully, these questions can guide you, and Auto Insights can help by automatically providing the answers you need to tell a great story.

If you want to start telling your data stories, but don’t have Auto Insights, check out this page for more info, a free trial, demos, use cases, and more! And if you already have Auto Insights, you can find training videos here.