Beyond the Bar Chart: Data into Actionable Insights

Raw Bar Charts? Really?

Ever seen someone explaining a raw bar chart in a presentation? If you did, please forward them this article and tell them to subscribe to
Analytics Wisdom!

Our inclination towards narratives makes storytelling and visualizations an effective tool for communicating complex data insights. The audience can better understand and remember the presented information by framing data insights in a narrative format. For example:

• A data analyst might use storytelling to explain the impact of a new marketing campaign on customer engagement by presenting the data as a story with a beginning, middle, and end.

• A data scientist might use a narrative approach to illustrate the results of complex statistical analysis by telling the story of how the data was collected, analyzed, and ultimately led to a particular conclusion.

• Product owners can leverage storytelling in data to simplify complex data and create a more engaging and memorable way of communicating insights to drive product development.

The 3-step Method to Plot Better Graphs

Example Case Study

Let’s take an example of a coffee chain that wants to display its customer’s median coffee bag purchases per month after the first purchase over 6 months. This data will be used to make strategic decisions for the business and even potentially raise funding. How does the coffee business approach this?

We will be using the seaborn package from python for this graph.

Step 1: Plot a Graph Suitable for the Data

It doesn’t always have to be a pie chart or bar graph, there are various ways to display data clearly and concisely. While bar graphs and pie charts are popular choices, there may be better options. Sometimes, a combination of different graph types, such as bar and line graphs, can effectively depict multiple metrics.

For example, a bar graph can show the total revenue for a specific product, while a line graph can show the trend over time. Heatmaps are another excellent option for cohort analysis, as they can highlight patterns and trends in large datasets.

Some common graph types and suitable situations :

• Bar Graphs: Comparing changes over time for categorical data (e.g. sales figures for different products over the past year)

• Line Graphs: Showing trends over time (e.g. stock prices over the past year)

• Scatterplots: Showing relationships between two variables (e.g. correlation between temperature and ice cream sales)

• Pie Charts: Showing relative proportions of different categories (e.g. market share of different companies)

• Heatmaps: Visualizing large datasets with multiple variables (e.g. cohort analysis for customer retention)

Step 2: Add Data & Axis Labels and Color Formatting

Displaying data best visually involves labeling data and color formatting the graph. In our example, if we have a bar graph that shows the Median coffee bags purchased after an initial purchase per customer for a 6-month time frame, we can add data labels to the bars to show the exact values. Additionally, we can label the X and Y axes to provide context and clarity to the reader.

We can then add color formatting to draw attention to the most important insights, making the data more visually appealing and easier to interpret. For instance, we color the bars here using a gradient from red to light green to green depending on the Median coffee bags purchased, with darker green indicating a higher number of coffee bags purchased. By using these techniques, we can make the data more accessible and easier to understand, leading to better decision-making and insights.

Step 3: Add Major Observations, Insights, and Takeaways

Step 3 involves adding essential details and insights that help convey the message of the graph. For instance, the graph mentioned in the previous paragraph has a title and subheading, a clear median line with a text box, a shaded area to highlight significant insight, and a text box with a question and answer. Adding such features to a graph provides context and helps the audience understand the data better.

Finally, it is essential to keep the graph simple, use appropriate colors and labels, and avoid cluttering the graph with unnecessary information. Best practice advice for adding significant observations, insights, and takeaway is to focus on the message you want to convey and provide supporting details that enhance the audience's understanding of the data.

Thank you for your time good ser 🫡 

References: 

• [https://medium.com/geekculture/techniques-of-data-visualization-from-charts-to-storytelling-a87850a1019b]

• [https://towardsdatascience.com/charts-that-tell-a-story-turn-a-plotly-visualization-into-something-more-a723e427d5aa]