Data Analysis tools
Open Extensions → Solvella → Data Analysis…, pick a tool on the left, fill in the form and press OK. By default the output goes on a new sheet named after the tool. Choose Output Range to put it somewhere specific. It won't write over cells that already have something in them unless you tick Overwrite.
Output is plain values, laid out the way Excel lays it out, so formulas you already have that point at Excel's ToolPak output usually only need the sheet name changed. Unlike Excel, nothing is a live formula: rerun the tool if your data changes.
If your first row or column holds headings, tick Labels.
| Tool | What it gives you |
|---|---|
| Anova: Single Factor | Group summaries and the one-way ANOVA table (SS, df, MS, F, P-value, F crit). |
| Anova: Two-Factor With Replication | Cell summaries and the two-way table with interaction. Set Rows per sample. |
| Anova: Two-Factor Without Replication | Row and column summaries and the two-way table. |
| Correlation | Pearson correlation matrix (lower triangle, like Excel). |
| Covariance | Covariance matrix. Note it's the population covariance (divides by n), which is what Excel does. |
| Descriptive Statistics | Mean, standard error, median, mode, SD, variance, kurtosis, skewness, range, min, max, sum, count, and optionally confidence level and k-th largest/smallest. |
| Exponential Smoothing | Smoothed series for a damping factor, with optional standard errors. |
| F-Test Two-Sample for Variances | Means, variances, F, one-tail P and F critical. |
| Fourier Analysis | FFT or inverse of a range whose length is a power of 2 (up to 4096). Results are complex numbers as text, like 3+2i, same as Excel, so IMABS and friends work on them. |
| Histogram | Bin counts, with optional cumulative %, and sorted (Pareto) output. Leave the bin range empty to have bins chosen for you. |
| Moving Average | Moving average for an interval, with optional standard errors. |
| Random Number Generation | Uniform, normal, Bernoulli, binomial, Poisson, patterned or discrete. Set a seed to get the same numbers again. |
| Rank and Percentile | Each value's rank and percentile, sorted from largest. |
| Regression | Regression statistics (R², adjusted R², standard error), the ANOVA table, and coefficients with standard errors, t, P and confidence limits. Columns that are exact combinations of others are dropped, as Excel does. Pro adds residuals, standardised residuals, probability output and confidence levels other than 95%. |
| Sampling | Every n-th value, or a random sample of a given size. |
| t-Test: Paired Two Sample for Means | Paired t-test with one- and two-tail P and critical values. |
| t-Test: Two-Sample Assuming Equal Variances | Pooled-variance t-test. |
| t-Test: Two-Sample Assuming Unequal Variances | Welch's t-test. Excel rounds the degrees of freedom to a whole number before working out P; we do the same so the numbers match. There's an option to use exact df instead. |
| z-Test: Two Sample for Means | z-test with known variances you type in. |
| Logistic Regression (Pro) | Not in Excel. For a 0/1 outcome: coefficients, standard errors, z, P, odds ratios with confidence limits, log-likelihood, likelihood-ratio test and McFadden's pseudo R². |