Make your data clear
Preparing your data is a critical step in effective data storytelling – it clarifies your metrics and reshapes your dataset so your charts reveal the full story.
Find the right metric for your story
The metric you choose can completely transform your data story, and it must directly serve your core message. Before selecting any metric, write out the single sentence you want your audience to remember.
For example:
- Is your message about growth? Choose percentage change.
- About disparities? Show differences from average.
- About market dominance? Display share of total.
Your core message should dictate your metric choice, not the other way around.
Check that your metric supports your message
The metric you choose should directly support your core message. If your key takeaway is that “Product launches are accelerating” but the chart shows total products per year, the metric doesn’t match the story – you would need to show the right unit in a chart type that makes the trend visible.
A useful check is to read your chart title alongside the visualization. If the metric doesn’t clearly support the claim in the title, consider calculating a derived metric or identifying additional data that better supports your message.
Commit to one metric (unless contrast is the message)
Once you’ve identified the right metric, use it consistently. Switching between totals, percentages, and growth rates can confuse audiences and weaken your message.
The exception is when contrast is part of the story – for example, “While absolute sales increased, our market share declined.” In this case, the difference between metrics helps highlight the insight.
Whenever you use derived metrics or calculations, document them clearly in your footer to maintain transparency.
The best metric for your story might not exist in your raw data.
Be prepared to derive new metrics through calculation (growth rates, differences from benchmarks) or gather additional data entirely (competitor figures, historical context) to truly illuminate your story.
Shape your data structure to support the chart
Datasets rarely arrive in the exact structure needed for a visualization. You may need to reshape, filter, or calculate new metrics before building your chart.
Different chart types also require data in different formats, so understanding your dataset helps ensure your chart shows the right relationships clearly.
Wide vs long data formats
One of the most common structural changes is converting between wide and long data formats.
Wide format (common in spreadsheets):
This format works well for charts that compare categories directly, such as bar charts or simple line charts.
| Month | Product A | Product B |
|---|---|---|
| Jan | 120 | 90 |
| Feb | 150 | 110 |
Long format
Many chart types (including scatter plots) require this structure because each observation must appear in its own row.
| Month | Product | Revenue |
|---|---|---|
| Jan | A | 120 |
| Jan | B | 90 |
| Feb | A | 150 |
| Feb | B | 110 |
Switch between formats with pivoting and unpivoting
Changing between wide and long formats is one of the most common data preparation tasks when building visualizations.
From wide to long format
In Flourish, you can quickly convert wide data into long format using the unpivot feature in the Data tab. This automatically turns multiple columns into a category column and a value column — allowing your dataset to work with chart types that require long-form data, such as scatter plots, small multiples, or faceted visualizations.
From long to wide format
If you need to go in the opposite direction — converting long data into wide format — tools like Excel or Google Sheets can help. Pivot tables allow you to reorganize rows into columns so your data can be used in charts that expect separate series.
These simple transformations make it easier to adapt your dataset to the chart type you want to use, without manually rewriting your data.
Common ways to derive new metrics
Sometimes the metric in your dataset doesn’t directly support the story you want to tell. In these cases, calculating a new measure can reveal the pattern more clearly.
- Growth rate: show how quickly something is changing over time.
Example: monthly revenue growth = (current month − previous month) ÷ previous month. - Share or percentage: show how categories contribute to a total.
Example: product revenue as a percentage of total sales. - Differences from a benchmark: compare performance against targets or previous periods.
Example: actual revenue vs forecast. - Per-unit metrics: normalize totals to make fair comparisons.
Example: revenue per customer or sales per region.
Restructuring your dataset and calculating the right metrics ensures your chart highlights the pattern you want your audience to see.
Numerical clarity and focus in Flourish
Flourish can transform unwieldy figures into clear, accessible insights without sacrificing accuracy or detail.
Order data to reveal patterns
Use Flourish’s sorting options to arrange your data in ways that support your message.
For bar charts comparing categories, sort by value (descending or ascending) to immediately show rankings and hierarchies.
Format numbers for readability
Use Flourish’s number formatting to convert raw data into clear information. Add prefixes ($, £, €) and suffixes (k, M, B, %) for context without clutter. Adjust decimal places according to the data’s precision needs.
Focus on meaningful axis details
Use Flourish’s tick controls to show only 4–6 key ticks. This helps viewers orient themselves without adding unnecessary detail. If your chart displays values directly on bars or data points, removing it can also create a cleaner visualization that keeps the focus on the data.
Linear or logarithmic axis?
Most charts use a linear axis, where equal distances represent equal numerical differences. This works best when values are in a similar range and you want viewers to compare absolute changes.
A logarithmic axis shows equal ratios instead of equal differences. This can be useful when values grow exponentially or span very large ranges.
The two charts below have the same data: can you spot the difference? The “traditional” linear scale shows the company’s meteoric rise – the famous hockey stick curve – around the time they launched the iPhone. This is a chart about acceleration and high growth. The chart with the log scale shows fluctuations in the company’s revenue in the early years, the steep rise in the early 2000s, and the slower growth in the last decade.
As a rule of thumb, use linear scales for most charts, and consider log scales only when large value ranges hide meaningful patterns.
Read more on our blog
