ARBIMETRA LAB
Covariance Analysis
Shows how variables move together in their original units, with sample and population matrices.
What it does
Shows how variables move together in their original units, with sample and population matrices.
Use it as an input to portfolio risk and multivariate methods.
Data you need
- Two or more numeric variables
Supported data: Numeric. Level: Beginner.
Formula and method
Cov(x,y) = Σ(x−x̄)(y−ȳ) / (n−1)
Each value is taken as a deviation from its mean, and pairs of deviations are multiplied and summed. Dividing by n − 1 gives the sample covariance, and by n the population covariance, across rows complete on all variables.
Assumptions
- Observations are paired
Input checks
- Listwise-complete rows
- Missing rows need confirmation
How to read the result
The sign gives direction. The size depends on units, so compare pairs using correlation.
Worked example
Illustrative run on Simulated study data (teaching data). These figures are the actual output of this model on that example, and they say nothing about your own data.
- The largest sample covariance in absolute size is 72.5876, between pre_score and post_score.
- pre_score and post_score tend to move in the same directions.
Limitations
- Magnitudes are not comparable across different units
Numerical validation
Validated 15 of 15 automated checks passed, including numerical results, invalid input, missing values, small samples, constant variables, charts and the interpretation text.
| Check | Reference | Result | Abs. error | Tolerance | Source |
|---|---|---|---|---|---|
| Anscombe sample covariance | 5.501 | 5.501 | 8.9e-16 | 1e-3 | Anscombe (1973) by calculation |
| Sample against NumPy | 5.501 | 5.501 | 0.0e+0 | 1e-12 | Independent NumPy calculation |
| Population against NumPy | 5.00090909 | 5.00090909 | 0.0e+0 | 1e-12 | Independent NumPy calculation |
Computed with python 3.13.2, numpy 2.2.5, scipy 1.14.1, pandas 2.3.3, statsmodels 0.14.4, scikit-learn 1.7.0, in your browser. A passing check confirms the listed comparison only. It does not make any result suitable for your data, and it is not a claim of equivalence with other statistical software.
Questions
Why is my covariance so large?
Covariance is in the product of the two variables' units, so large units give large values. Correlation removes the units.
ARBIMETRA LAB is an educational and research tool. Results should be checked by the researcher and are not professional advice.
