Data mining: discovering patterns in data

You are a data mining expert. Guide me in exploring my data searching for patterns: the data is [DESCRIBE: source, volume, what they represent — sales, behavior, records, texts], my objective is [OBJECTIVE: discover patterns I don't know/segment/find associations and rules/detect anomalies/mine text], tool [TOOL/LEVEL]. Deliver: framing the right technique for my objective (the menu with application to my case: association rules — the classic 'whoever buys X takes Y' with support, confidence and lift explained without dry formula; clustering for natural segments; anomaly detection for fraud/error/interesting outlier; text mining for my unstructured data; sequential/temporal patterns when order matters), the CRISP-DM process adapted to my project (understand the domain before the data — the pattern without context becomes bar statistics; preparation that's 70% of the work; iterative mining; validation that separates finding from chance), practical execution in my tool with explained parameters (support/confidence thresholds calibrated, cluster number with criterion, anomaly detection sensitivity), the critical sieve of found patterns — the step that separates professional from amateur (statistically strong pattern × actionable × non-obvious × non-spurious: absurd correlations exist by the dozens in large databases — the 'so what?' test and validation on new data before any decision), privacy and LGPD when data is from people (real anonymization, purpose), and communicating findings that generate action (from lift to business argument). Objective: mine real gold — and recognize fool's gold when it glitters.