Visual Analytics: Choosing the Right Chart
Seven visualization techniques and the analytical situations where each one works best.
Effective data visualization is not simply about making data visually appealing. The choice of chart should reflect the analytical question being asked.
This collection presents seven of my preferred visualization techniques in Tableau. Each visualization is designed for a different analytical purpose—from explaining how individual components contribute to a total, to tracking changes in rank, mapping geographic patterns, analyzing flows, identifying dense intersections, and comparing hierarchical composition.
Rather than treating these charts as interchangeable design options, this project focuses on when each visualization is most useful, why it works, and when a simpler alternative may communicate the data more effectively.
The Selection Principle
Choose the chart based on the question—not the data type alone.
- Contribution → Waterfall
- Ranking over time → Bump Chart
- Geographic patterns → Spatial Map
- Part-to-whole comparison → Stacked Bar
- Flows and transitions → Sankey
- Two-dimensional intensity → Heatmap
- Hierarchical composition → Treemap
01 — Waterfall Chart
Analytical Question: How do individual components contribute to the movement from a starting value to a final result?
A waterfall chart decomposes a cumulative result into the positive and negative contributions that create it. Instead of showing only the final total, it reveals how each component moves the running value upward or downward.
Best Used For
Profit bridges; revenue-to-profit analysis; budget variance; cost decomposition; period-over-period financial changes; explaining how individual categories contribute to a total.
Why It Works
The sequential structure makes contribution visible. Viewers can follow how each component changes the cumulative result rather than comparing isolated values.
When Not to Use It
Avoid waterfall charts when the primary goal is simply comparing independent categories. A standard bar chart is usually clearer when cumulative contribution is not part of the analytical question.
02 — Bump Chart
Analytical Question: How does the relative ranking of categories change over time?
A bump chart emphasizes movement in rank rather than absolute values. Each line represents a category, allowing changes in competitive position to be followed across multiple periods.
Best Used For
Market-share rankings; product rankings over time; sports standings; regional performance rankings; competitive position analysis.
Why It Works
Connecting ranks across periods makes overtaking, decline, stability, and changes in relative position immediately visible.
When Not to Use It
Do not use a bump chart when exact magnitude matters more than ranking. A line chart is generally more appropriate for analyzing actual values over time.
03 — Spatial Map
Analytical Question: Where are observations occurring, and are there meaningful geographic patterns?
Spatial maps connect measures and categories to geographic location, making regional concentration, geographic variation, and spatial relationships visible.
Best Used For
Store or facility locations; healthcare access; transportation analysis; regional sales; service coverage; geographic clustering; market expansion analysis.
Why It Works
Location itself becomes an analytical dimension. Patterns that may be difficult to recognize in a table become intuitive when placed in geographic context.
When Not to Use It
Avoid maps when geography is incidental to the question. If precise comparison between regions is the primary objective, a ranked bar chart may communicate differences more accurately.
04 — Vertical Stacked Bar
Analytical Question: How is a total divided among categories, and how does that composition compare across groups or periods?
A stacked bar combines total magnitude with categorical composition. Each bar represents a total while its segments show how individual categories contribute to that total.
Best Used For
Revenue composition; customer-segment mix; product-category contribution; survey-response composition; channel mix; comparing part-to-whole relationships across periods.
Why It Works
It allows viewers to examine both the overall size of a group and its internal composition in a single view.
When Not to Use It
Avoid heavily segmented stacked bars when exact comparison between interior segments is important. Grouped bars or small multiples are often better for precise category comparisons.
05 — Sankey Chart
Analytical Question: How do observations flow from one category or stage to another?
A Sankey chart visualizes movement between categories. The width of each connection represents the magnitude of the flow, making both dominant pathways and smaller transitions visible.
Best Used For
Customer journeys; conversion funnels; patient pathways; referral patterns; traffic-source transitions; supply chains; movement between states or categories.
Why It Works
Sankey charts preserve the relationship between origin and destination while simultaneously communicating the relative magnitude of each pathway.
When Not to Use It
Avoid Sankey diagrams when there are too many nodes or connections. Excessive flows quickly create visual clutter. For simple stage totals, a funnel or bar chart may be clearer.
06 — Heatmap
Analytical Question: Where are the strongest and weakest intersections between two dimensions?
A heatmap uses color intensity to encode magnitude across a matrix of categories. It is particularly effective for revealing concentrations, gaps, clusters, and outliers across two dimensions.
Best Used For
Day × hour activity; product × region performance; correlation matrices; website engagement patterns; scheduling and demand analysis; operational intensity patterns.
Why It Works
Color intensity enables rapid scanning across many combinations simultaneously, allowing patterns to emerge without requiring every value to be read individually.
When Not to Use It
Avoid heatmaps when users need highly precise individual values or when the dataset contains only a small number of comparisons. A table or bar chart may then be easier to interpret.
07 — Treemap
Analytical Question: Which categories dominate a total, particularly when the data contains many categories or hierarchical groups?
A treemap represents categories as nested rectangles whose area corresponds to magnitude. It allows many categories to be displayed compactly while preserving part-to-whole relationships.
Best Used For
Portfolio composition; product-category contribution; market segmentation; budget allocation; hierarchical datasets; identifying dominant categories.
Why It Works
Treemaps use space efficiently and make disproportionately large or small categories visually apparent.
When Not to Use It
Avoid treemaps when precise comparison between similarly sized categories is important. Humans compare length more accurately than area, so a sorted bar chart is preferable for exact comparisons.
Visualization Is a Decision
The same dataset can often support several technically valid charts, but technically valid does not necessarily mean analytically useful. Effective visualization starts by identifying the decision, comparison, or pattern the audience needs to understand and then selecting the visual structure that communicates it with the least unnecessary complexity.
The seven examples in this collection represent different analytical tasks rather than simply different visual styles.
- If the question is “What created this total?” → Waterfall
- “Who moved up or down?” → Bump Chart
- “Where is it happening?” → Spatial Map
- “What makes up each total?” → Stacked Bar
- “Where does it flow?” → Sankey
- “Where are the concentrations?” → Heatmap
- “What dominates the whole?” → Treemap