The analysis workbench for sensor time-series data
You have a file of readings from a logger, a weather station or a sensor you wired up yourself. Drop it in here and get the averages, the connections and the figures out of it — without installing anything, making an account, or writing a line of code.
It never leaves your browser. The button opens the resample & aggregate tool, which is where most people start — every other tool reads the same file.
13 of the 13 tools below work today. The rest are listed anyway and are marked as not built yet. Nothing here is a mock-up: if a page says a tool works, the arithmetic runs.
Your file stays on your machine
There is nowhere to upload it to. The site is static files and the analysis is JavaScript in your tab — load the page, pull the network cable out, and keep working. How to check that.
It reads logger files as they come
Onset HOBOware and Postlad exports are recognised on sight, including the awkward variants — no plot title, no row numbers, tab-separated, the date and the time in separate columns. You should not have to tidy a file up first.
It tells you what it hid
Averaging throws information away, so every aggregation here reports the highest and lowest reading it averaged over, and a gap in your data is drawn as a gap rather than as a line straight across it.
What you can do with it
Every tool opens with a real dataset already loaded, so you can see exactly what it produces before you go and find your own file.
Clean up and average
A reading every five minutes is too much to look at. Turn it into hourly or daily numbers — and see what the averaging hid.
Turn a logger CSV at any sampling interval into hourly, daily or monthly values — mean, minimum, maximum, sum or median — with the within-bucket range preserved.
Find connections
Which of your channels move together, and which one moves first. Answering the second question is the reason this product exists.
A Pearson or Spearman correlation matrix across every channel in a logger file, with the coefficient and the number of readings behind it printed in every cell.
Compute the cross-correlation of two time series across a range of lags and read off the lag at which they line up best.
Weather and growing
The standard agricultural and meteorological quantities. Every one of them worked out across your whole record, not from one typed-in reading, and every formula cited on the figure.
Dew point
FreeTurn a temperature and humidity record into a dew-point series, with the dew-point depression and the spells where the air reaches saturation.
VPD calculator
FreeCompute vapour pressure deficit across a whole temperature and humidity record, air and leaf, with the saturation vapour pressure equation stated and cited.
Accumulate growing degree days from your own temperature record by the simple-average or single-sine method, with a base temperature and an optional upper cutoff.
ET₀ (FAO-56)
FreeCompute daily reference evapotranspiration by the FAO-56 Penman-Monteith method from a weather-station export, not from one set of typed-in numbers.
Accumulate winter chill from an hourly temperature series by the 32–45 °F, Utah and Dynamic models, and compare what the three of them say.
Trend tests for reports
The statistics a monitoring report has to contain, done the way the guidance says — including what to do with values below the detection limit.
Run the Mann-Kendall trend test with tie handling and Sen’s slope with confidence limits on a monitoring series, including non-detects, without installing anything.
Find every gap, stuck sensor and implausible jump in a logging record, and get the completeness figure that belongs in a methods section — per channel and per calendar month.
Everything else
Also in the catalogue.
Run the Mann-Kendall trend test and Sen’s slope across every well and every analyte in a monitoring dataset at once, and read the answers off one report-ready matrix.
Track the correlation between two sensor channels over a sliding window and see where it changed, against the whole-record coefficient and the spread sampling alone would produce.
Count the exceedance events in a sensor record and total the time spent past a threshold — per day, month or season — with the time inside gaps reported as unknown rather than counted as compliant.
Not sure which one?
Start with Resample & aggregate. It reads your file, tells you what it found in it — how many readings, how often, which time zone, where the gaps are — and turns it into hourly or daily values. Once you can see the shape of your own data, the rest of the list stops looking abstract.
Open resample & aggregateIs the arithmetic right?
Worked examples from FAO-56, Gilbert, the EPA and NIST, put beside what this site computes — with the test that asserts each one named underneath it.
Will it read my logger file?
What each export actually looks like, and which of its oddities are handled: HOBO, Campbell TOA5, METER ZENTRA, Lascar EasyLog.
I already use something else.
What the other thing is better at, and the one job it does not set out to do: Grafana, ThingSpeak, or desktop trend-testing software.
Where the readings come from
Plotlad works on files. If you are still collecting readings by walking out to a logger and plugging it into a laptop, its companion — Postlad, at postlad.com — takes an HTTP request from an Arduino or an ESP32 and gives you a live chart and a share link, with no account. Two separate products at two separate prices; the link between them is a URL, not a subscription.
Coming laterOpening a live Postlad stream straight into a tool here, without exporting a file first. That does not exist yet — today it is export, then drop.
Pricing here will be a one-time $49 for twelve months or $119 for a founding licence. Neither is on sale yet, and everything marked free stays free.