| name | lemma-eda |
| description | EDA kickoff for a fresh dataset: fixed opening scaffold (goal, imports, load, sanity), then chapters derived from the data; scan leakage, land a baseline. |
| homepage | https://github.com/tkpratardan/lemma |
| license | MIT |
Exploratory data analysis
Deliver
Provide a trustworthy dataset orientation, a short ranked set of findings,
important limitations, and the next analytical question or baseline suggested
by the evidence.
Check
- Establish observation unit, source identity, coverage, candidate identifiers
or targets, and a seed for sampled work.
- Inspect shape, types, ranges, duplicates, missingness, time coverage, and
group rates.
scripts/profile_table.py is available for a deterministic
first pass.
- Explore only patterns that could change a decision or next step.
- Investigate suspicious quality, leakage, subgroup, or temporal patterns
before recommending modeling.
- State a conclusion for each explored branch.
Do not generate an indiscriminate chart catalog, silently drop missing rows,
infer causality, or turn a bounded question into open-ended EDA.
For visualization and chapter patterns, read
references/deep-guide.md.
