Time-series slope over a range: the rate of change in a window

Many series have a phase that matters — the linear growth phase, the release between two time points, the warming after a switch — and the question is the rate in that phase. MinutePlot fits an ordinary least-squares line to the points inside a window you choose and reports the slope with its standard error, R² and p, for every series in the table.

The data

Time in the first column, one column per series (two cultures here):

Series layout: x in the first column, each further column one series; blank cells are simply skipped.

In MinutePlot

Choose Time-series slope over a range, paste the table, set the window (here 4–10 h, the exponential-looking stretch) and run. The figure shows every series with the window shaded and the fitted line drawn through it; the table gives slope, SE, R², p and n per series. The endpoint slope and the area under the curve are available as alternatives to the fitted line.

The result

Culture A: 0.156 per hour (R² = 0.99, p < 0.001); culture B: 0.108 per hour (R² = 0.99, p < 0.001). Both rates are precisely estimated, and A grows about 45 % faster in the window.

The figure MinutePlot produces: both series, the chosen window, and the fitted line through each series inside it.

Styling and exporting this figure: see Figure styling and Exporting & downloading.

Report it like this

“Between 4 and 10 h, optical density rose at 0.156 h⁻¹ in culture A and 0.108 h⁻¹ in culture B (linear fits, R² > 0.99, p < 0.001).”