Power BI Lesson 41 – Visualization Types | Dataplexa

Visualisation & Service · Lesson 41

Visualisation Types

Every chart type encodes data in a specific way — position, length, angle, area, colour saturation. Choosing the right chart for the right data is not aesthetic preference; it is about which encoding lets the audience make the most accurate comparisons with the least cognitive effort. This lesson covers every visualisation type in Power BI, when to use each one, when to avoid it, and the specific configuration decisions that determine whether a chart communicates or confuses.

The Visualisation Decision Framework

Before picking a chart type, answer three questions: What type of data am I showing? What relationship am I showing? Who is the audience and how much time do they have? The answers narrow the field considerably.

Relationship to show	Best chart types	Avoid
Change over time	Line chart, Area chart, Column chart (for fewer periods)	Pie chart, Donut, Scatter (unless showing trajectory)
Part-of-whole	Donut chart (≤5 slices), Stacked bar/column, Treemap (for many items)	Pie chart with >5 slices, 3D pie (distorts angles)
Ranking / comparison	Bar chart (horizontal), Column chart (vertical for time), KPI visual	Line chart (implies continuity), Donut/Pie
Correlation / distribution	Scatter chart, Bubble chart	Bar chart (hides the relationship), Line chart
Geographic	Map, Filled Map, Azure Maps, Shape Map	Bar chart unless geography is not the point
Exact values / detail	Table, Matrix, Card, Multi-row card	Any chart — if you need exact numbers, use a table

Bar and Column Charts

Bar charts (horizontal) and column charts (vertical) encode values as length — the most accurately perceived visual encoding after position. They are the workhorse of business reporting and the correct default choice for almost any comparison between categories.

Clustered Bar — horizontal, categories on Y axis

Laptop Pro

48K

Desk Chair

22K

Office Desk

18K

Mouse

9.5K

Use when: category labels are long, ranking is the point, or there are many categories (>6). Horizontal allows more label space.

Clustered Column — vertical, time on X axis

Jan

Feb

Mar

Apr

May

Jun

Use when: time is on the X axis with ≤12 periods, or comparing two series side by side within a time period. Vertical bars feel more "timely" than horizontal.

Variant	When to use	When to avoid
Clustered	Comparing two to three series side by side — e.g. this year vs last year by month	More than three series — bars become too thin to read
Stacked	Showing total while revealing how parts contribute — e.g. revenue by category per month	When comparing the inner segments is important — the floating baseline makes comparison hard
100% Stacked	Showing percentage composition over time — e.g. category share of revenue by month	When absolute values matter — you cannot read magnitude from a 100% chart

Line and Area Charts

Line charts imply continuity — the line between two data points suggests a trend. This makes them ideal for time series where the in-between values are meaningful, and wrong for categorical data where the "line" between "North" and "South" has no interpretation.

Line chart — trend over time

Use for: ≥7 time periods, continuous trends, YTD vs prior YTD comparison, two or more measures on the same axis.

Area chart — volume under a trend

Use for: emphasising cumulative volume (YTD revenue). Avoid stacked areas for more than two series — the floating baselines make comparison impossible.

Pie and Donut Charts

Pie and donut charts encode values as angles and arc lengths — both less accurately perceived than bar length. Research consistently shows that people misjudge slice sizes by 10–30%. Use them only when the message is "one segment dominates" and there are five or fewer categories. If you need precise comparison, use a bar chart instead.

✓ Good use of donut chart

Three segments, one dominant: Electronics 62%, Furniture 25%, Accessories 13%. The dominance of Electronics is immediately clear. The centre label reinforces the key figure.

❌ Bad use of pie chart

Eight product categories, all between 8% and 18%. None dominates. The viewer cannot accurately compare 11% vs 13% from arc length alone. A ranked bar chart communicates the same data in half the time with twice the accuracy.

Rule: if your first instinct is to add a data label to every slice, use a bar chart instead.

Card and KPI Visuals

Card visuals show a single number prominently — the most space-efficient way to communicate a key figure. The KPI visual adds a comparison value and a trend indicator. Both are essential on executive dashboards where the primary question is "are we on track?" not "exactly how much was each region?"

Total Revenue

$110K

Card visual — single measure, large display

Revenue YoY %

+21.4%

▲ vs 52,000 prior year

KPI Visual

$110K

▲ +21.4% vs target

The KPI visual requires three fields: Value (the measure), Trend axis (usually Calendar[Date] or a month column), and Target (a goal measure). The trend line is drawn automatically. The indicator arrow and colour are driven by whether Value exceeds Target.

Table and Matrix

Tables show exact values and are appropriate when the audience needs to look up specific numbers, export data, or see detail that cannot be aggregated. A matrix adds row and column groupings — the equivalent of a pivot table — and supports subtotals, conditional formatting, and drill-down.

Matrix — Revenue by Region (rows) and Quarter (columns)

Region	Q1	Q2	Q3	Q4	Total
North	15,900	14,200	16,800	18,100	65,000
South	9,200	8,800	10,100	11,300	39,400
West	5,600	6,200	6,800	7,400	26,000
Total	30,700	29,200	33,700	36,800	130,400

Scatter and Bubble Charts

Scatter charts plot two measures against each other to reveal correlation. Each dot is one entity — a customer, a product, a store. A bubble chart adds a third measure as the bubble size. Both are underused in business reporting despite being the best way to answer "which products have high revenue but low margin?" or "which customers are high value but high churn risk?"

Bubble chart — Revenue vs Margin %, bubble size = Order Count

Bubble size = Order Count · Upper-right quadrant (Stars) = high revenue AND high margin · Lower-right (Cash Cows) = high revenue but low margin — focus pricing review here

Treemap and Decomposition Tree

Treemap — part-of-whole with many items

Laptop Pro
43.6%

Chair 20%

Desk 16%

Mouse 8.6%

Others

Use when: many categories, part-of-whole relationship, size differences are large. Avoid when categories are similar in size — the rectangles become indistinguishable.

Decomposition Tree — drill-down attribution

Revenue 110K

→

Electronics 62K

Furniture 40K

Accessories 8K

→

North 25K

West 22K

South 15K

Use when: explaining what is driving a variance — "why did revenue drop this month?" Users click through multiple dimension levels to find the root cause.

Waterfall Chart

The waterfall chart shows how an initial value changes through a series of positive and negative contributions to reach a final value. It is the standard chart for financial variance analysis — "we started with $100K budget, spent $40K, received $15K reimbursement, and ended with $75K."

Waterfall — Revenue bridge: actual vs budget variance by category

Green = positive contribution · Red = negative · Blue = start and end totals · Essential for budget vs actual bridges and P&L analysis

Gauge and Funnel Charts

Gauge chart

Shows a single value on a min-to-max arc, with a target needle. Instantly communicates "how close to target." Requires three values: minimum, maximum, and the target. Use for operational metrics with a clear target — customer satisfaction score, uptime %, defect rate.

Avoid when: the audience needs to compare multiple gauges simultaneously — the arc encoding is slow to compare across multiple visuals. Use a bullet chart (custom visual) instead.

Funnel chart

Shows sequential stages of a process where values decrease at each step — conversion rates, sales pipeline stages, onboarding steps. Each segment's width represents the value at that stage. The visual immediately shows where the biggest drop-off occurs.

Avoid when: the stages are not a true sequential funnel — using a funnel for unrelated categories misleads the audience into seeing a process that does not exist.

Teacher's Note: The most common visualisation mistake in Power BI is not choosing the wrong chart type — it is using too many chart types on one page. A dashboard with a line chart, a donut, a scatter, a treemap, a gauge, and a waterfall all on one page creates visual noise that prevents any single insight from landing. The principle to remember is: one page, one question. Every visual on a page should contribute to answering one clearly defined business question. If you find yourself adding a sixth visual type to a page, that is usually a signal to create a second page rather than a more complex first one.

Practice

Practice 1 of 3

You need to show revenue trends across 18 months and highlight that the overall direction is upward. The chart type that implies continuity between data points — making it ideal for time series with many periods — is the ___ chart.

Practice 2 of 3

A donut chart with eight slices all between 8% and 18% is difficult to interpret because humans perceive ___ less accurately than bar length, making it hard to compare similarly-sized slices.

Practice 3 of 3

The ___ chart type is best suited to financial variance analysis — showing how an initial value (e.g. budget) changes through a series of positive and negative contributions to arrive at an actual final value.

Lesson Quiz

Quiz 1 of 3

A manager asks for a chart showing which products have high revenue but low margin — to identify pricing opportunities. Which chart type best answers this question and why?

A stacked bar chart with Revenue and Margin % as two series per product — the side-by-side bars directly show which products have high revenue and low margin in the same visual space A scatter chart with Revenue on the X axis and Margin % on the Y axis, one dot per product. Products in the upper-right quadrant (high revenue, high margin) are stars. Products in the lower-right quadrant (high revenue, low margin) are the pricing opportunity targets. No other chart type reveals this two-dimensional relationship as clearly. A treemap with Revenue as the size and Margin % as the colour saturation — the largest tiles represent high-revenue products and darker tiles show low margin, revealing the combination at a glance A ranked bar chart sorted by Revenue with a secondary line showing Margin % on a dual axis — the visual comparison between the bar height and line position reveals which high-revenue products have below-average margins

Quiz 2 of 3

A line chart shows monthly revenue for North, South, and West regions across 12 months. The West line is barely visible and appears flat near the bottom. What is the most likely cause and how do you fix it without changing the chart type?

The Y axis minimum is set to zero, which compresses the West line relative to the much larger North and South values. Fix: change the Y axis minimum to a value closer to the lowest West data point — this is appropriate when comparing the rate of change rather than the absolute magnitude, but should be clearly labelled to avoid misleading the audience about the true difference in scale. The West line is filtered out by a visual-level filter — check the Filters pane for the visual and verify all three regions are included in the Region filter The West data uses a different measure than North and South — check the Values field well and confirm all three series reference the same Total Revenue measure Line charts cannot display more than two series accurately — add a small multiples configuration to give West its own panel with an independent Y axis scale

Quiz 3 of 3

An executive dashboard has eight visuals on one page: a line chart, a donut, a scatter, a treemap, a waterfall, a gauge, a matrix, and a KPI card. A design review flags the page as too complex. What is the most principled fix?

Reduce the font size on all visuals to make them smaller and fit the page more efficiently — smaller visuals leave more white space which reduces perceived complexity Remove the scatter and treemap charts as they are the least common Power BI chart types and most executives are unfamiliar with them — replace them with additional bar charts to reduce cognitive load Define one clear business question this page answers and remove every visual that does not directly contribute to answering it. Move the remaining visuals to separate pages — each page addresses one question. A well-designed executive summary should have three to five visuals maximum, all serving the same question. Use bookmarks to hide half the visuals by default and show them on demand — this keeps all eight visuals available without presenting them simultaneously, reducing initial complexity while preserving all the data

Next up — Lesson 42 covers Visual Formatting in depth, including the Format pane structure, consistent colour themes, data label configuration, axis formatting, and the specific settings that determine whether a report looks professional or amateurish.

← Previous Course Index Next →

Power BI Course