Document missing-value decisions and manual review tables

README.md

@@ -17,6 +17,8 @@ configs:
   data_files: data_dictionary/data_dictionary_summary.csv
 - config_name: missing_values_summary
   data_files: data_dictionary/missing_values_summary.csv
+- config_name: missing_and_leakage_manual_review
+  data_files: data_dictionary/missing_and_leakage_manual_review.csv
 ---
 
 ![LendingClub Loan Data banner](assets/banner.svg)
@@ -48,11 +50,19 @@ This dataset backs **Credit Risk Default**: an explainable credit-risk platform
 
 ## Data Dictionary
 
-Two project-generated summary files under `data_dictionary/`, produced while cleaning
-the raw LendingClub CSV (see [`notebooks/00_cleaning.ipynb`](https://github.com/BuildersLab/Credit-Risk-Default/blob/main/notebooks/00_cleaning.ipynb)), not Kaggle's raw column list. They have different columns, so each is declared as its own config above (see `configs:` in the metadata) rather than being auto-merged.
+Reference files under `data_dictionary/`, produced while cleaning the raw LendingClub CSV (see [`notebooks/00_cleaning.ipynb`](https://github.com/BuildersLab/Credit-Risk-Default/blob/main/notebooks/00_cleaning.ipynb)), not Kaggle's raw column list. Each has a different schema, so each CSV is declared as its own config above (see `configs:` in the metadata) rather than being auto-merged.
 
 - **`data_dictionary_summary.csv`**: one row per column: `Variable`, `Description`
   (LendingClub's official column definition), `Missing Values` (count),
   `Percentage (%)`, and `Missingness Label` (e.g. "Very High (95-100%)").
 - **`missing_values_summary.csv`**: missing-value counts and percentages per column,
   used to identify and drop columns with structural missingness before modeling.
+- **`missing_and_leakage_manual_review.csv`**: the full manual review of all 151
+  original columns, `Keep` (boolean), `Reason`, and whether each column is
+  available at time of application (pre-loan) versus only during the loan
+  lifecycle (post-origination). This is the source review behind the leakage
+  and structural-missingness decisions in `00_cleaning.ipynb`.
+- **`missing_values_decisions.xlsx`**: per-column imputation decision for the 87
+  columns that survive the leakage/keep review (median fill, median plus a
+  missing-flag column, mode plus flag, or drop rows), with the mechanism and
+  reasoning behind each call.