A statistical sample is a subset of a larger data set selected for early inspection or analysis. In an ML project, it provides a practical first view of data before full extraction, helping reveal scale, quality, limitations, and possible solution directions. section 2.4
Create a data-discovery ticket that records:
- Statistical properties of the sample. section 2.4
- Nonfunctional properties, including scale, speed, and history. section 2.4
- System properties, including location, infrastructure, and operational purpose. section 2.4
- Known deficiencies, cleaning needs, and possible supplementary sources. section 2.4
Early data discovery
When extracting the full data set is technically difficult or currently unaffordable, inspect a sample first. Use the findings to validate known data sets, uncover overlooked sources, and estimate the work needed to clean or extend limited data. section 2.4
