Analyse statistics with AI
Choose a summary that suits your data, investigate unusual observations and judge whether a measured difference matters in practice.
To analyse statistics with AI, begin with the decision the figures are meant to support. A rise in average dispatch time might affect nearly every order, or a handful of exceptionally late parcels might account for most of it. The statistical-analysis guide covers summaries, variation, trends and tests that help distinguish those possibilities.
Keep the comparison periods and units visible. A week with a public holiday can contain a different mix of orders from a normal week. A difference may also be measurable yet too small to justify changing a shift. The size of the effect belongs beside the uncertainty around it.
What it does
- Compare the mean and median of a skewed measure
- Describe variation with percentiles or spread
- Investigate the influence of unusual observations
- Check weekly patterns before interpreting a trend
- Discuss the practical importance of a test result
Install and use
- Install the Data plugin in a compatible agent. The background guide applies to relevant requests. Supply an authorised dataset and a suitable Python environment for calculations.
- Ask “Compare dispatch times across these two periods. Show the median and spread, then explain how the longest delays affect the mean.”
- Check each group’s sample size and express the effect in the original unit before deciding what to change.
FAQ
Source
Made by Anthropic. Original skill: statistical-analysis, published on GitHub under the Apache-2.0 licence.
View on GitHub