Completeness asks whether recorded fields contain usable values. Incomplete observations may be mostly blank, or a dataset may contain seasons with too few records for reliable analysis. section 2.3.2 section 12.4
Handle incomplete data by:
- Inspecting summaries and filtered records to locate blank fields or unusually low record counts. section 12.4
- Removing mostly blank observations when imputation or another correction would not be useful. section 2.3.2
- Excluding incomplete seasons before creating a working dataset for further analysis. section 12.4
NBA salary data
The gini dataset contains only three Boston Celtics records for 1987 and four Los Angeles Lakers records for 1990. Removing seasons 1985–1990 reduces the data from 14,163 to 12,929 records, producing a more usable working dataset. section 12.4
draft <- draft[-c(90, 131),]
View the R expression that removes selected incomplete observations from a dataset. section 2.3.2
