Change over time

Time-series chart with event annotations

Adds dated events or shaded periods to explain the context around a time series.

What it answers

What happened around a notable event?

The question

How did weekly sign-ups respond to the campaign and the price change?

Sign-ups run near 215 a week, jump to 410 after the campaign starts in week 6, and drop back to about 235 after the price rise in week 11.

Illustrative data. It belongs to no company.

Adding launches, outages or policy changes as context.

When annotations overwhelm the data or imply unsupported causal claims.

Read the trend and locate dated annotations without assuming causation.

Coincidence with an event does not establish causation.

Layout: Two linked tables
Fields: Time: date; Value: number; separate EventTime and EventLabel records
One row is: One time-series observation; each event record describes a separate annotation. What is the grain?
Rules: Keep measured observations separate from annotation records. Use the same time zone.
Optional: EventEnd for shaded intervals. Do not copy an event onto every observation.
Preparation: Parse and sort timestamps or sequence positions; aggregate duplicates deliberately and preserve gaps. Keep measured observations separate from annotation records. Use the same time zone.
Units: Time uses a stated date/time unit and sampling period; the value axis uses the measure's unit. Rates need their own explicit denominator.

Two linked tables, joined by a shared key column.

Jan 1=30, Feb 1=34; launch Jan 15. Mark the launch. The later value is 4 higher; the chart cannot attribute that rise to launch.