Data cleaning & exploratory analysis
Start with CSV, Excel, or database results. Check missing values, duplicates, anomalies, and definitions, then run code to produce statistics and reviewable charts.
Inputs
Data files or read-only query results, field definitions, business questions, and date range.
Deliverables
Cleaning scripts, quality report, exploratory charts, findings, and analytical limitations.
Step-by-step workflow
Break the work into verifiable stages, then choose models and supporting tools.
- 01
Understand granularity and definitions
Identify what each row represents, its key, and definitions of dates and amounts.
- 02
Check data quality
Run checks for missing values, duplicates, anomalies, types, and join problems.
- 03
Explore distributions and relationships
Compute distributions and group differences, generate charts, and retain filter conditions.
- 04
Deliver reviewable results
Preserve source data, scripts, and processing logs. Connect conclusions to actual calculations.
Required model capabilities
Combine models for the actual stages. The directory includes candidates matching one or more of these capabilities.
View matching modelsSupporting tools
Selection & delivery checks
- Code generation must be paired with real execution; a written analysis alone is insufficient.
- Explain how missing-value handling and outlier removal affect results.
- Correlation does not establish causation.
An example request
Clean CSV data with Python, check missing and duplicate records, compute grouped metrics, create charts, and summarize findings and limitations.Find models for this request