Calculate
Statistics
Reads one figure off records or numbers — average, lowest, closing, a rolling twelve months. You pick the question; the data shapes the answer.
| Port | Side | What it carries |
|---|---|---|
in | In | Records, or several numbers — not both at once. |
window | In | An optional number of months for the rolling sum and average, such as a slider from 3 to 24. |
out | Out | The figure; for some questions, the row it came from or a series as well. |
What you choose
The question, from the menu on the card. Each entry says what it does.
| Question | What it gives |
|---|---|
| highest, lowest | The largest or smallest single value. |
| average, median | The mean, or the middle value. |
| opening, closing | The value before the first row, or after the last. |
| rolling sum | Each point totals the trailing months — twelve months to date at 12. |
| rolling average | The same window, divided by its length. |
| reset balance | The balance walked again from zero at the first row. |
| year-over-year | Each point against the same point a year back. |
| count periods | How long the data spans, in days, weeks, months, quarters or years. |
| count records | How many rows there are. |
| count series | How many different values a column holds — customers, currencies. |
| find patterns | Groups rows by who the money moved with, and measures how regular each is. |
With records wired in, you also choose the column on the card — amount, balance, or any other number the records carry. The rolling window is a field on the card; wire a number into window and it takes over for as long as the wire stands.
What Fluidly decides
- Records or numbers. Wire several numbers and the menu offers only what makes sense for them — count numbers in place of count records, and no questions that need dates.
- The row behind the figure. highest, lowest, opening and closing point at a real row. The output carries that row as well as the figure, so a Struct component can tell you the date it falls on.
- A series where there is one. Rolling sums, rolling averages and year-over-year send out the whole series alongside the latest figure — wire it into a Timeline to draw it.
- Patterns from the data. find patterns keeps every row and adds how far each sits from its group's usual amount and day. The filters downstream set the tolerance. For invoices, the card also asks which dates to compare — payment date against due date shows who pays late.
The name follows the question: Lowest, Rolling sum, YoY balance — and that name travels on to the Timeline's legend.
Example
Forecast → Statistics (lowest) → Range trigger → Email: an email the day the lowest balance the forecast expects drops below your line.