Data Science: What is it?

The work includes choosing what to measure, checking data quality, exploring patterns and explaining uncertainty. A useful outcome can be a corrected metric or a better question rather than a predictive model. Domain knowledge determines whether a technically correct calculation answers a meaningful question.

The NIST statistical handbook identifies John Tukey’s Exploratory Data Analysis as a seminal contribution. Exploring structure, anomalies and assumptions remains important even when computational tools change. Data science draws on older statistical and scientific traditions; it did not appear fully formed with today’s job titles.