Data analysis / Resquire guides
Ask a question. Follow the evidence.
Turn source files into an analysis your team can inspect, challenge, and use.
Useful analysis starts with a decision and a shared definition of the numbers. Give an agent the question and source material, then ask it to show how it reached the answer.
Define the question before the chart.
Use a channel where the people responsible for the decision can participate. Share the source files or accessible documents, and explain which metric and time period matter.
Write down definitions that could change the answer: what counts as an active customer, how refunds are handled, or which timezone determines the end of a day.
- Accessible source data
- Metric definitions and a time range
- The decision this analysis should inform
Ask for the working, too.
An answer is easier to trust when the calculations and assumptions are available. Ask the agent to distinguish observations from possible explanations.
Using the attached exports, compare weekly activation for the last eight weeks. Show your definition, calculations, and missing-data checks. Include a chart and the supporting table. Separate what the data shows from explanations we still need to investigate.
Check a few rows together.
Follow the files and tool activity as the analysis develops. If the agent is missing a source or a definition, resolve that gap in the conversation.
Compare a few results with the original data. Check denominators, filters, and date boundaries before discussing the recommendation. Ask the agent to rerun the analysis when an assumption changes.
Finish with a decision people can own.
Ask for a concise result with sources, calculations, caveats, and proposed next steps. Keep the supporting files beside the summary so another teammate can inspect them later.
Assign the business decision to a person. An agent can prepare evidence; the team decides how to act on it.