Record the inputs
Name the data source, the collection date, the unit of observation, and any exclusions. If you cannot share the data, explain that restriction and provide a safe representative example when possible.
Describe the method
Keep the transformation steps, tool versions, and parameters alongside the result. For a simulation, record the random seed. For a comparison, define the baseline and report the same metrics for each option.
Share uncertainty
Report limitations and alternative explanations. Distinguish a visual pattern from a statistical conclusion. Make clear which findings are exploratory and which hypotheses were specified in advance.
Make the next run possible
Give readers a short path to reproduce the result: where the inputs live, which command to run, and what output to expect. A small inspectable experiment is a better starting point than an impressive result nobody can verify.