Paste up to ten numerical columns and instantly compare their shape, spread, and central tendency without assuming any underlying distribution.
The Distribution Visualiser builds Kernel Density Estimation (KDE) curves for each column you supply, overlaying them on a common axis. KDE is a non-parametric method: unlike fitting a single named distribution, it makes no prior assumption about whether the data are normal, log-normal, or Weibull, which makes it the right first look when you are exploring data or comparing populations whose form you do not yet know.
A smooth KDE curve exposes structure a histogram can hide. Multiple peaks (multimodality) often signal that two failure mechanisms or two sub-populations are mixed in one dataset. Differences in width show changes in variability, and asymmetry shows skew that summary statistics like the mean conceal. The shared axis makes side-by-side comparison of vendors, batches, or operating regimes immediate.
Seal run-life from two suppliers is pasted as two columns. Supplier A's KDE is clearly bimodal, with one cluster near 90 days and a second near 320 days, indicating two distinct failure modes within the same part number rather than one consistent population. Supplier B's curve is unimodal and tight around 280 days. The mean run-life of the two suppliers is nearly identical, so a spreadsheet average would have called them equivalent; the density curves show A carries a hidden early-failure population that B does not.
For: engineers running statistical process control, vendor comparisons, or exploratory analysis of run-life and test populations.