Explore data structure, missing values and distributions to find useful leads.
Understand an unfamiliar dataset before deciding how to clean, analyze or model it. The skill checks field types, missingness, distributions and outlier clues in CSV, TSV and JSON, including group or time splits that could affect later analysis.
Provide the files, field background and the questions you want to investigate. It helps produce a reviewable exploration report describing the scope, main findings and useful next steps while preserving the source data. A second analysis path can be used for cross-checking where available. Specialist formats such as genomics or medical imaging have reference guidance rather than built-in automatic parsing; statistical inference and causal questions require further analysis.
Examples describe the request and the shape of the expected deliverable; they are not completed run results.
trial.csv is inside my authorized folder: confirm the format is supported, then show structure and missingness profiles, a leakage check grouped by condition over subject_id, identifiers kept anonymized.
A profile and leakage report with per-column missing/type notes; an unsupported format gets routed onward instead of force-read.
Compare the measurement column's distribution under log-transform versus winsorizing — bounded reports only, don't touch the raw file, and list proposed treatments for me to decide.
Side-by-side statistics under both treatments, described plots, a boundary-sample list — and the data unchanged.
iOS · Android
iOS
Android
MIT
MIT (root LICENSE.md)
MIT at the pinned upstream revision; full license and per-file adaptation/asset audit are preserved in references/licenses/. Floe additions and synthetic fixtures are identified separately.
Publication commit 25e07dbeeb, audited source 84c524b6fb; catalog digest iOS 46196ba57ed7… · Android 16de89d019e2….
Version records ship with the signed catalog; the target and released versions are shown below.
