Point chart

Who it’s for

Anyone looking at readings that arrive irregularly or should not be joined up — spot samples from a manual test, lab results, a sensor that reports on change rather than on a schedule. Drawing a line between such readings invents data between them; points do not.

What it does

The Point chart plots each reading as a discrete marker on a time axis, with no line connecting them. Two series render as two sets of 8 px circles, so gaps in reporting stay visible as gaps rather than being spanned by a straight segment that implies readings nobody took.

• Gaps stay honest — a device that reported twice today shows two points, not a line across the silence.
• Each point is one reading, which makes clustering, scatter and outliers directly visible.
• Point shape and size are per series, so overlapping sets stay distinguishable without relying only on color.
• Point labels can print the value beside each marker for charts with few enough readings to carry them.
• Several Y axes, thresholds, comparison overlays and data zoom all work as they do on the other time series charts.
• The Point chart is a preset of the Time series chart with lines turned off and points on: turning the line back on for one key is a per-key setting rather than a different widget.

How to set up

Data keys

The Point chart takes one set of points per data key, from one or more datasources.

Key	Role	Type	Description
First	One set of plotted points	Timeseries	Timeseries only. Each key’s settings carry its point shape, size and labels, and which Y axis it uses.

Use a raw or lightly aggregated time window — heavy aggregation collapses the individual readings this chart exists to show, and the result is a sparse line chart drawn without its line.

How to customize

• To separate overlapping series — give each key a different point shape as well as a different color.
• To make sparse readings easier to hit — raise the point size, which also makes a chart with few readings look deliberate rather than empty.
• To print values beside the markers — turn on point label for that key, and set its position, font and color.
• To join some readings after all — turn the line back on for that key, which is useful when one continuous series provides context for scattered ones.
• To stop one series flattening another — add a Y axis with the second series’ units and assign the key to it.
• To mark a limit — add a Threshold from a constant, an attribute or another key.
• To let readers inspect a period — leave Data zoom on.

Tips

• Turn on point labels whenever the chart holds fewer than about thirty readings. Sparse data is usually sparse because each reading was expensive to obtain, and the exact values are then worth more than the trend.
• Keep the time window aligned to how often the source reports. A window ten times longer than the reporting interval leaves the points crowded at one edge, and the emptiness reads as an outage rather than as an oversized window.