
Published: 15 November 2026 · Updated: 15 November 2026
Paste your data and get the full summary instantly: mean, median, mode, quartiles, range, and both sample and population standard deviation — plus least-squares regression with correlation and predictions for paired data. It starts with 2, 4, 4, 5, 7, 9, 10 (mean 5.8571, sample SD 2.9114). Free, no sign-up.
Paste data, read the summary. Mean, median, mode, quartiles and both standard deviations update as you type, and the regression box fits ŷ = a + bx to paired data with r, r² and a prediction at any x.
Researched by MixKhazana Team · Last updated 15 November 2026 · Data: OpenStax Statistics §12.2; OpenStax Introductory Business Statistics §2.7
A statistics calculator turns a list of numbers into its summary: the centre (mean, median, mode), the spread (range, quartiles, standard deviation), and — for paired data — the straight line that best fits the cloud of points. Statistics courses expect students to compute these by hand once, then use a calculator for real data sets, where hand calculation is slow and error-prone. The College Board expects a graphing calculator with statistical capabilities for AP Statistics. This free tool covers the same descriptive statistics and linear regression for homework and revision.
The standard deviation is the heart of descriptive statistics: a single number that says how far typical values sit from the mean. Two classes can share the same average score yet be completely different — one tightly bunched, one wildly spread — and only the spread measures reveal that.
Let the data be x1, …, xn, with mean x̄ = (sum of all values) / n:
The sample formulas divide by n − 1 rather than n because the sample variance estimates the population variance: dividing by n − 1 gives the better estimate, a point every statistics textbook stresses. The regression line is the line that minimises the sum of squared errors (the SSE), which is why it is called the least-squares line.
Find the mean and sample standard deviation of 2, 4, 4, 5, 7, 9, 10:
Try it above: the default data give exactly these numbers. For regression, the default pairs (1,2), (2,4), (3,5), (4,4), (5,5) give slope 0.6, intercept 2.2 and r = 0.7746.
| x | x − x̄ | (x − x̄)² |
| 2 | −3.8571 | 14.8776 |
| 4 | −1.8571 | 3.4490 |
| 4 | −1.8571 | 3.4490 |
| 5 | −0.8571 | 0.7347 |
| 7 | 1.1429 | 1.3061 |
| 9 | 3.1429 | 9.8776 |
| 10 | 4.1429 | 17.1633 |
| Total | 0.0000 | 50.8571 |
The deviations always add to zero — that is why they must be squared before averaging. Squaring also punishes far-out values: the single value 10 contributes more than a third of the total.
The slope b says how much ŷ changes, on average, for a one-unit increase in x: in the default example, 0.6. The intercept a is the predicted y when x = 0, which often has no real-world meaning — it is just where the line crosses the axis. The correlation r measures the strength and direction of the straight-line relationship: 0.7746 is a fairly strong positive association. And r² = 0.6 means 60% of the variation in y is explained by the line. None of this proves causation: a strong r between two variables never, by itself, shows that one causes the other.
The sample version divides the squared deviations by n - 1, and the population version divides by N. Use the sample version when your data are a sample from a larger group, which is the usual case in class. Use the population version only when you have measured every member of the group.
Because the sample variance is an estimate of the population variance, and dividing by n - 1 gives the better estimate. Textbooks call this Bessel's correction in spirit: with n - 1 the average of many sample variances equals the population variance.
The mean adds all values and divides by the count; the median is the middle value when the data are sorted. The median resists outliers: one huge value pulls the mean up but barely moves the median.
r measures the strength and direction of a straight-line relationship, from -1 (perfect falling line) through 0 (no straight-line relationship) to +1 (perfect rising line). It says nothing about causation.
It is the line y-hat = a + bx that makes the sum of squared vertical distances from the points to the line as small as possible. Any other line has a larger sum of squared errors.
Predictions inside the observed x range are interpolations and are the most trustworthy. Predictions far outside it are extrapolations: the straight-line pattern may not continue. Also check r-squared: a low value means the line explains little of the variation.
Linear interpolation on the sorted data (the common spreadsheet percentile method): the position is (n - 1) x p, interpolating between neighbours. Different textbooks use slightly different quartile rules, so small differences from your textbook are normal.
Yes. Copy a column from a spreadsheet and paste it into the data box: values separated by new lines, commas, spaces or semicolons are all accepted.
Disclaimer: a study aid for homework and revision. Quartiles use linear interpolation; textbooks differ slightly in quartile rules, so small differences are normal. It is not an approved exam-room device.
Card photos: Sergey Meshkov / Pexels, Pexels licence · Yulia Meshkova / Wikimedia Commons, CC BY 4.0 · Joost J. Bakker / Wikimedia Commons, CC BY 2.0 · Coyau / Wikimedia Commons, CC BY-SA 3.0