Need a second set of eyes? Get human help →

Outlier Finder

Detect anomalies that create work anxiety — spikes, duplicates, and suspicious rows you can investigate before anyone else does.

Input
CSV / TSV / XLSX / XLS / delimiter-style TXT
Loaded:none
Safety: your file is read locally in your browser. Use Masker if you need to share samples.
Works best with numeric columns (sales, cost, hours, qty, metrics). Everything runs locally.
Run summary
How many anomalies were found
Rows: Cols: Numeric cols: Outliers: 0 Rules hit: 0
Tip: Use a time column + group column (like Store, SKU) to detect spikes within each group.
1) Choose columns
Pick a value column to score, plus optional grouping & time
If group is set, the tool detects outliers within each group (ex: per store, per customer, per SKU).
2) Detection method
Different methods catch different anomaly shapes
Sensitivity 3.0
Higher sensitivity = fewer, stronger outliers. Lower sensitivity = more findings.
ignore unusually small values
prevents false spikes from missing values
3) Quality filters
Reduce noise, improve signal
helps merge duplicate groups like "Store 1" vs "Store 1 "
limits output to top N scores
If your dataset is huge, consider exporting just the top 100–500 anomalies first.
Preview
Top outliers with scores + reasons
Audit
Method details + thresholds
Load a file, pick a value column, then run outlier detection.
Back to Toolbox Contact