A live stem-and-leaf plot: every arriving value flies in from a staging slot and drops into its tens-stem row, existing leaves shoulder-nudging aside to insert it in sorted order, so the distribution is built entirely from the digits of the data and every leaf is still an individually focusable record.
Use when the distribution matters but so does every individual record, since each leaf is that value's own digit and stays a focusable table cell rather than dissolving into a bar; pick histogram-live-grain instead when the sample is large enough that individual records stop being readable and a settling heap is the better read, or chart-bar-halftone when the buckets are fixed categories rather than derived from the data's own digits.
checkout · response time stem-and-leaf plot. 42 values, median 132.5ms, range 79ms to 231ms. Each row is a stem of ten ms; each leaf digit is one value's ones digit.
70–79 ms · 1 value
80–89 ms · 2 values
90–99 ms · 5 values
100–109 ms · 2 values
110–119 ms · 4 values
120–129 ms · 5 values
130–139 ms · 8 values
140–149 ms · 1 value
150–159 ms · 5 values
160–169 ms · 0 values
—
170–179 ms · 0 values
—
180–189 ms · 0 values
—
190–199 ms · 3 values
200–209 ms · 1 value
210–219 ms · 2 values
220–229 ms · 2 values
230–239 ms · 1 value
n 42median 132.5ms · range 79–231msstem = tens
rolling window · 60 samples
every arrival types its ones digit into the row for its tens — hover or focus any leaf for the record behind it
Props
Prop
Type
Default
Description
records
StemSiftRecord[]
—
rolling window of raw samples, oldest first
unit?
string
—
unit suffix for readouts and stem labels, e.g. "ms"