Data Viz for Dummies
A guide to choosing a chart
Start with the question you want to answer.
A chart choice starts with a question: who leads, what changed, how spread out are the values, or where did the flow go? This guide uses one fictional sports season to introduce the main chart families. Read the question above each example, compare the marks, and notice what the form makes easy to see. You do not need to know the component names to begin.
Taxonomy · the pregame show
Do not ask ‘which chart?’ Ask what kind of truth is trying to get out.
The same data can suggest several charts. Sort this guide by the kind of data you have or by the task you want to do. Some chart families appear in both groups because they support more than one task.
Use the map to narrow your choices, then check whether the data supports the comparison you want to make.
Interactive chart family map
Organized by kind of data
Shared lanes are visual cousins: a line and scatterplot both welcome paired values, while bars and treemaps can both host categories—but with radically different reading costs.Bar chart · point guard
Behold the bar: clean-cut nepo baby of position, still earning the minutes.
Bars compare values by length from a common baseline. Sorting them makes the ranking easy to read and lets you compare neighboring players at a glance.
Start the value axis at zero when bar length represents the total. A cropped baseline can make small differences look much larger.
Player leaderboard · points per game
M-Train leads the room; the order does the explaining
Read from top to bottom for rank, then compare adjacent bar lengths for the size of each gap.Grouped bar · wing defender
Comparison is ranking after it acquires subplots, lore, and a shared calendar.
Grouped bars put each player’s measures together. Compare shot locations within a player, or follow the same color across players. A small number of series is easiest to keep track of.
When there are too many groups or colors to follow, try separate small charts, a dot plot or a filter for the comparison that matters.
Scoring profile · points by zone
Same totals, different basketball biographies
Zuri lives in the paint; Leo bends the perimeter. Grouping preserves both the player total and the shot-location composition.Line chart · transition offense
Time enters the gym and suddenly every point has a before and an after.
A line connects observations in order. It makes changes and trends easy to follow, which is useful when comparing performance across a season or any other sequence of measurements.
Connect values only when their order matters. Use straight segments or clearly explain smoothing so readers can distinguish observations from the path drawn between them.
Ten-game arc · two team signals
The season found a second gear in week six
Points and team assists rise together after week five. The annotation directs attention to that change.Histogram · sixth player
The average arrives late, says ‘my bad,’ and misses the shape of the whole party.
A histogram groups values into ranges and counts how many fall in each. It shows where possession lengths cluster and how widely they vary. Try different bin widths to check whether a pattern depends on the grouping.
Read the distribution as a whole. Include the sample size and bin widths, and inspect the observations before explaining an unexpected peak.
Possession clock · 32 trips
Most possessions resolve in the middle; a long tail burns clock
Most possessions fall around 14–18 seconds, with a few longer ones stretching the right side of the distribution.Scatterplot · two-way playmaker
Two variables meet. Correlation enters the chat. Causation remains on read.
A scatterplot compares two measurements for each observation. Look for clusters, trends and outliers. Bubble size can add a third measure when that extra comparison is useful.
A trend shows how values vary together. To explain why, you will need to consider other factors and how the observations were selected.
Minutes × assists · bubble = rebounds
Minutes create opportunity; role shapes what happens next
Leo’s assist total outruns the minutes pattern, while Omar’s large rebound bubble occupies a different job description.Sankey · floor general
A total becomes a journey: where it came from, where it went, what got lost.
A Sankey follows quantities through stages. Ribbon width shows how much takes each route, making it useful for budgets, journeys and possessions. Use consistent units so readers can follow the amounts from start to finish.
If too many ribbons cross, filter or group the routes. Explain any gains or losses between stages so the totals remain understandable.
100 possessions · outcome paths
The paint is productive; resets are expensive
Follow the thickest ribbons. Early pushes and half-court sets both favor paint touches, and those touches convert more often than resets.Treemap · roster depth
Hierarchy is a house of rooms; area tells you which rooms consume the rent.
A treemap fits a hierarchy into nested rectangles. Larger rectangles represent larger values, making it easy to locate the main branches. Labels and nesting explain how each part belongs to the whole.
For an exact ranking, use bars. For parent-child relationships with little emphasis on values, use a tree. A treemap is useful when both structure and relative size matter.
Playbook · usage share
Attack owns the binder; every rectangle keeps its family name
Area estimates use frequency; nesting reveals strategic family. The chart helps locate dominant plays, not read a box score to the decimal.Choosing a chart
Name the task, then choose the form.
Complete the sentence “I need my reader to…” and use the tasks below to choose a starting point.
Find the leader
Sorted barCommon baseline. Zero drama. Elite scan game.
Compare subgroups
Grouped barKeep the legend short or facet the roster.
See what changed
LineTime or another defensible order belongs on x.
Understand spread
HistogramInterrogate the bins; report the sample.
Test association
ScatterplotShow the points. Causality is not in the contract.
Trace movement
SankeyThe ribbon width must conserve one quantity.
Browse nested parts
TreemapOverview first; exact comparison second.
Say one number
BigNumberAdd context, target, date, and what changed.
A useful chart makes the comparison you care about easier to see.