Check a CSV before modeling
Point the Skill at an approved CSV in a dedicated folder. Review detected column types and missingness in the redacted summary, then ask the data owner to verify any ambiguous field before selecting features.
Exploratory Data Analysis
Use the Exploratory Data Analysis Skill to profile approved local data before modeling. Its core tools summarize supported files without network access and flag limits the reviewer needs to see.
Inspect a local dataset before analysis
Record the allowed input and check aggregate structure and missingness before choosing an analysis.
Provide the approved folder, filename, and the decision this inspection will inform.
Copy the request into a new Floatboat conversation and provide the required material.
Download FloatboatYour request for Floatboat
Preview and copy the request. It is not sent automatically from this page.
Reference: Exploratory Data Analysis https://github.com/K-Dense-AI/scientific-agent-skills/blob/main/skills/exploratory-data-analysis/SKILL.md Read the original documentation and check the current environment and dependencies. Propose any new installation or tool connection before proceeding. My task: Use K-Dense AI’s exploratory-data-analysis Skill to profile my approved local CSV with the bundled core script. Show the file and allowed directory before running. Report aggregate missingness, sampling limits and follow-up checks without exporting rows or modifying the source. Expected deliverable: A redacted aggregate report with explicit coverage and data-quality checks.
About Exploratory Data Analysis
K-Dense AI’s MIT-licensed Skill supplies bounded Python inspectors and analysis instructions. Its standard-library core handles CSV, TSV and strict JSON with redacted aggregate reporting. Optional formats have separate dependencies. Use the results to decide what to verify before modeling, and keep unsupported files outside the inspection.
Exploratory Data Analysis Use Cases
Use the same boundaries when comparing files, but vary what you inspect according to the next research decision. Keep the original data and its aggregate review separate.
Compare two TSV deliveries
Profile both approved exports with comparable limits. Compare their column counts and aggregate missingness, noting sampling and schema differences before asking for a corrected delivery.
Inspect a strict JSON export
Provide a local JSON file and the expected record structure. Use the bounded inspection to identify unexpected structure, then record which checks need an approved follow-up.
Review a possible leakage field
Provide the prediction target and data split rules alongside the approved data. Review candidate leakage signals separately from confirmed defects and ask for source evidence before removing a field.
How to Run a Bounded Data Inspection
Choose an explicit local directory and supported format. Review the aggregate profile and its coverage before relying on any suggested data-quality finding.
Specify the approved inputs
Name the allowed local root and list the exact files. Keep inputs outside the task scope out of the run.
Run the supported core inspector
Use the Skill’s bundled script for the file type with Python 3.11 or later and preserve the aggregate output.
Review coverage and limitations
Compare the report with the file type, limits and next analysis decision. Keep the source files intact.
Explore More Agent Skills
Read another Skill guide and check its original source before using it.
Exploratory Data Analysis FAQ
What file types does the core support?
The standard-library core supports CSV, TSV and strict JSON. Optional inspectors have their own format and dependency requirements.
Can the Skill inspect data without network access?
The core scripts run locally without network access. Check the separate dependency notes before choosing an optional inspector.
Will the report show individual data records?
The upstream workflow calls for redacted aggregate results and excludes raw rows, direct identifiers, and full paths.
What if the file format is unknown?
Leave it unprocessed. Confirm the format and use a documented inspector or arrange a separate approved workflow.
Can I use the report as a statistical conclusion?
Use the report to identify questions and quality checks. Confirmatory analysis requires a separate method and validation plan.
Which license applies to this Skill?
The specific upstream SKILL.md lists MIT. Check other manifests individually before assuming the same terms.