Essential session storage is required to sign in. Optional analytics remain off unless you allow them.
Take a real, messy public dataset (e.g. a Kaggle CSV, a city open-data export, or survey results) and produce a clean, analysis-ready table plus an EDA writeup. Document the cleaning decisions: handle missing values (drop vs impute, justified per column), remove duplicates on a real key, detect outliers with the 1.5*IQR fence and decide fix-vs-exclude, and standardize formats/types. Then explore: summary statistics, distributions, and at least two relationships, with honest charts. End with 3-5 findings stated plainly.
Missing/duplicate/outlier/format issues handled and JUSTIFIED; outliers via 1.5*IQR with fix-vs-exclude reasoned; validation checks run.
Correct summary stats (mean vs median chosen appropriately for skew), distributions examined, relationships explored without overclaiming causation.
Right chart per question; axes/scales not misleading; clear labels.
3-5 clearly stated, data-supported findings; limitations noted (sampling, missingness, correlation != causation).
Use this README structure to make the work reviewable.
DevPath fetches a read-only, size-capped snapshot. Submitted code is never executed.
# Clean and explore a messy real-world dataset ## Goal Take a real, messy public dataset (e.g. a Kaggle CSV, a city open-data export, or survey results) and produce a clean, analysis-ready table plus an EDA writeup. Document the cleaning decisions: handle missing values (drop vs impute, justified per column), remove duplicates on a real key, detect outliers with the 1.5*IQR fence and decide fix-vs-exclude, and standardize formats/types. Then explore: summary statistics, distributions, and at least two relationships, with honest charts. End with 3-5 findings stated plainly. ## Acceptance criteria - [ ] Cleaning rigor: Missing/duplicate/outlier/format issues handled and JUSTIFIED; outliers via 1.5*IQR with fix-vs-exclude reasoned; validation checks run. - [ ] EDA quality: Correct summary stats (mean vs median chosen appropriately for skew), distributions examined, relationships explored without overclaiming causation. - [ ] Honest visualization: Right chart per question; axes/scales not misleading; clear labels. - [ ] Findings: 3-5 clearly stated, data-supported findings; limitations noted (sampling, missingness, correlation != causation). ## Verification - [ ] Document installation and run commands. - [ ] Record automated checks and their exact commands. - [ ] Add representative output, screenshots, or a short demo where useful. - [ ] Confirm failure paths and known constraints. ## Decision log ### Decision title - Context: - Choice: - Alternatives considered: - Tradeoffs: ## Evidence - Link each rubric criterion to the file, test, or artifact that demonstrates it. ## Known limitations - List what is intentionally out of scope and what should be improved next.