data-profiler

To profile a dataset before analysis, you need a systematic protocol that discovers column inventory, defines the unit of observation, checks identifier uniqueness, and flags missingness patterns. This process outputs a structured Markdown profile document detailing variable types and data quality signals.

Checking panel data balance and merge feasibility requires evaluating identifier uniqueness, panel entry/exit patterns, and logical inconsistencies. A dedicated dataset profiling protocol tests these empirical finance-oriented checks against common research sources to ensure your keys align properly.

Generating a variable dictionary and missingness report involves profiling variables to identify types and calculating missingness rates. Using R with tidyverse conventions, you can load CSV files to automatically discover structure and output a structured Markdown dataset profile.

Yes, you can profile large datasets using R with tidyverse conventions, utilizing data.table specifically for large datasets. The profiling protocol discovers file structure and defines the unit of observation to identify potential panel balance issues before analysis.

Dataset profiling does not automatically save files; it requires user approval gates before writing. It produces a structured Markdown dataset profile using a provided template and recommends a save location, ensuring you confirm before any files are written.

You should not use an automated dataset profiling protocol when your workflow requires write operations or data transformations, as this protocol is designed for read-only exploration. It is strictly limited to discovering structure and flagging data quality issues without modifying the source data.