Humans are natural storytellers. Storytelling has always been a primary means for transmitting knowledge across groups of people, resulting in the formation of cultures and enabling evolutionary success. And now as we enter the digital age where data is the primary currency, we turn to data storytelling yet again to communicate complex information, using compelling narratives customized for specific audiences to increase engagement and user adoption.
Understanding data storytelling
Data in itself cannot resonate with humans as effectively as stories can. With stories, you get a distinct start, middle, and ending. They have a narrative flow and repetition; they tie information together, providing a framework that the audience can easily follow. While data describes what is happening, stories indicate why it is happening.
That said, it takes specific skills and tools to effectively communicate findings to others and convince them to take action.
What are the skills required to be a good data storyteller?
To be an effective data storyteller, you need to have analytical, UI/UX, data visualization and soft skills such as creativity, communication, and the ability to craft a narrative that will be meaningful to your audience.
- Analytical skills include knowledge of BI tools, data analysis, information architecture, and development methodology.
- UI/UX skills encompasses expertise in cognitive science, design thinking, color theory, user interface design, prototyping, and information graphics.
- Data visualization skills covers visual storytelling, visual problem framework and knowledge of various charts.
Best practices to create effective data stories:
Here is a summary which talks about six best practices for building effective data stories:
1. Tailor the story to the audience
There are broadly three types of audiences – strategic, analytical, and operational. Data stories change as per the type of audience.
Strategic
Strategic data storytelling is for decision-makers and senior management. They provide high-level performance measures with snapshots of weekly, daily, and monthly data.
The best practices for a strategic dashboard are as follows:
• Avoid using advanced visualizations and too many details
• Highlight outliers.
• Focus on actionable insights instead of making the dashboard visually attractive.
• Use thresholds and highlight negative and positive values.
Analytical
Analytical insights and applications are for mid-management and planning teams. They deliver complex data with rich comparisons. Interactive displays and historical data are included.
Here are the best practices for an analytical dashboard:
• Highlight insights correctly.
• Use interactive visualizations.
• Focus on actionable insights instead of making an attractive dashboard.
Operational
Meant for operational teams, this type of data stories monitors constantly changing activities and shows near real-time or real-time data.
To create effective operational dashboards, be sure to:
• Keep simple but actionable visualizations.
• Avoid placing too many details.
• Make the most of pre-attentive attributes.
• Focus on actionable insights instead of the dashboard’s aesthetics.
• Use thresholds and highlight the positive and negative values.
2. Understand the anatomy of a dashboard
The dashboard is crucial in data storytelling as it visually displays critical information needed to attain one or more objectives. Ideally, it should fit on a single computer screen to make it viewable at a glance. InfoCepts recommends including these elements on your dashboard:
- Corporate branding
- Dashboard header
- Date and currency
- Report area
- Report header
- Footer
- Summary section
If you have multiple dashboards, enable layout navigation so users can switch between them. It is also essential to include primary and secondary navigation menus, information icons, menu icons, and a call to action.
3. Learn how to make the most of color palettes
Colors can make your data storytelling more interesting and easier to follow. Color palettes can be:
- Sequential (ordered from low to high)
- Diverging (two sequential colors with a neutral midpoint)
- Categorical (contrasting colors for individual comparison)
- Alert (a color to convey a warning), or
- Highlight (a color for showcasing a particular value).
Consider these factors when choosing color palettes:
- Color code: Avoid using a combination of green and red in the same display as color-blind people cannot distinguish color-coded groups of data.
- Use of thresholds: Use red and green to represent negative and positive values respectively.
- Printing: Distinguish the colors for printing. Most printers print in black and white, so consider using contrasting colors.
- Avoid flashy colors and a bright background: Use a background color that sufficiently contrasts with the objects in the graph or table for optimum visibility.
4. Implement visual design best practices
- Ensure visibility on the entire screen –
Users may overlook critical information if they scroll to access data, so ensure all information is visible in one go. - Present information in a hierarchical manner –
Place the summary information at the top left corner, highlight important information, and organize related information groups.
- Effectively highlight important information -Direct the viewer’s eye to the most critical information first.
- Avoid showing excessive details –
Dashboards must have high-level information for a quick overview. As such, too many precisely expressed details can slow down the viewers. - Use visualization appropriately. –
Successful data storytelling must involve the apt representation of information but avoid overusing the visualization. Likewise, avoid oversizing visualization or stretching the visuals to occupy white space. - Avoid data ink elements –
Do not use dark background colors for graphs and grids, avoid bright fluorescent colors, use grid lines when necessary, and skip using borders for legends.
- Limit font types –
Too many font types make the text harder to read, and the font can look squished.
- Format your grids-Headers must stand out from the body. Delete non-data ink and keep attribute headers and values left-aligned. Align metric headers and values to the right and ensure a clearly visible total.
5. Pick the suitable chart
- Use a bar chart, deviation bar chart, or dual axis bar chart , especially when showing comparisons, patterns, or relationships.
- Use a stacked bar chart for patterns, part-to-a-whole, relationships, proportions, and comparisons.
- Use a line chart for displaying quantitative values over a continuous period or intervals, such as when showing relationships or trends.
- Use a pie chart to show percentages and proportions between categories.
- Use a bubble chart to compare and show relationships between categorized or labeled circles with proportions and positioning. It can provide an overall picture for analyzing correlations and patterns.
- Use a scatter plot to identify relationships existing between different values.
- Use an area chart to show the development of quantitative values over a period or an interval, making it ideal for displaying relationships and trends.
- Use a box plot to display numerical data groups through their quartiles. It can show the data distribution based on percentiles, minimum, median, and maximum, making it ideal for descriptive statistics and comparing distributions.
- Use a Gantt chart or project timelines for planning and monitoring resource allocation or project development on a horizontal time scale.
- Use a HiLow stock or candlestick chart for financial data to represent high, low, closing, and opening values.
- Use a histogram to visualize data distribution over a given time or continuous interval.
- Use a Pareto chart to identify the cause of a loss or problem. A histogram shows the frequency of a problem or the various problems occurring.
- Use a polar or radar chart to compare multiple quantitative variables. It is effective for viewing variables with similar values or showing if there are some outliers. It is also practical for highlighting high- or low-scoring variables in a dataset, making it suitable for showing performance.
6. Use advanced visualizations where needed
Here is a quick overview of a few advanced visualizations and when to use them in your data storytelling:
- Waterfall –
This visualization highlights decrements and increments of the values of metrics over time. It may be used for what-if analyses.
- Gauge –
Gauges are status indicators similar to a speedometer. They display the value of a single metric, making them practical as visual representations of a single metric value for benchmarks and comparisons.
- Time series –
An area graph that lets a document analyst determine a section of the graph to view at a time to explore a high-level trend or more metrics. It is suitable for data over time, patterns, and comparisons.
- Maps –
A map lets users visualize data to identify and analyze trends, patterns, and relationships in the data.
- Heat map –
A map consisting of colored rectangles representing an attribute element, so you can quickly grasp the impact and state of a large number of valuables. Heat maps are helpful for data storytellingin the financial services industry.
- Bubble grid –
This visualization makes it easy to identify prime trends or data anomalies relative to the total contribution of the accompanying data. It is ideal for conducting analyses involving key business ratios.
- Funnel: Use this to analyze various trends quickly over several metric values.
- Graph matrix –
This is an interactive and powerful visualization. It allows users to show their data using various graph styles (e.g. grid, bubble graph, or line). It can be customized to the user’s needs, allowing the comparison of metric trends by two non-time attributes.
- Lipstick chart –
This visualization offers different ways of viewing data. For every x-axis value, a gray bar shows the actual recorded value. Red and green bars represent the expected value.
- Micro charts – Use a micro chart for data storytelling that needs to show the metric’s trend at a glance without additional details. It can be a bullet, sparkline, or bar micro chart.
- Network visualization – Use this to easily and quickly identify relationships between related clusters and items. It is best for market basket analysis or visualizing a social network.
- Image layout – This lets users show the data as map markers or colored geographical regions. Display options can be changed, such as the color of the regions, allowing users to grasp relationships between different locations.
- Data cloud – This visualization shows attribute elements in various sizes to indicate differences in metric values between them. An element with a bigger font means a higher metric value.
- Media – Use this when you need to present information through channels like images, video, website content, or audio on the dashboard. Media can also enhance a dashboard’s look and feel.
With so much data available within an organization, data storytelling offers the best way to put a human perspective into the increasingly complex and rapidly changing world of digital disruption. Learn more tips on becoming an effective data storyteller from the full version of the InfoCepts Guide to Becoming a Data Storyteller.