Chart Annotation

Marking the thing that happened, so a reader is told what the shape means.

Also known as: annotating charts, chart captions, callouts on charts

Annotation is the text you put on a chart to say what a shape means: a label beside the line, a shaded band over the release window, an arrow at the dip, a rule at the target. A chart shows that something changed; annotation says what changed, so the reader does not have to guess or ask.

Why it matters

The person who made the chart knows why the line jumped. Everyone else sees a jump and supplies their own reason — usually the one that suits them. An unannotated spike in a time series turns into “the campaign worked” or “quality dropped” depending on who is talking, and neither claim is supported by anything on the page.

What to annotate

The events that explain a shape: a deploy, a price change, an outage, a campaign start, a public holiday, a change in how the metric is measured.
The thing you want noticed: the crossing point, the gap widening, the level the number used to sit at.
What the reader should do about it — even if the answer is “nothing, this is expected”.

Practical rules

Put the label next to the mark it describes, not in a legend at the edge.
Keep the words short. An annotation competes with the data for attention; if the caption is longer than the chart, the chart has become the decoration.
Date the events. “Outage” is not useful six months later; “outage, 14–16 Mar” is.
One annotation layer, not five. If everything is marked, nothing is.

The trade-off

Annotation is interpretation, and it should be attributable. “Expected dip — marketing pause” is a fact about a decision. “Growth slowing” is a claim the chart has not made. Keep what happened separable from what you think it means, and the chart stays usable by people who disagree with you. That separation is most of what data storytelling is about, and the discipline that keeps annotation from becoming decoration is the one in chart junk.