added a DataCleaner

src/data_cleaner.py

@@ -0,0 +1,90 @@
+import pandas as pd
+import numpy as np
+
+class DataCleaner:
+    """
+    General-purpose cleaner for multiple WORC datasets
+    (Employment, Enrollments, Demographics).
+
+    Uses try/except for safety (does not break if col missing).
+    Keeps all rows (no drops), but fills/fixes when possible.
+    """
+
+    def __init__(self, df: pd.DataFrame):
+        self.df = df.copy()
+
+    def safe_drop_columns(self, cols_to_drop):
+        """Drop columns if they exist, otherwise ignore."""
+        try:
+            self.df = self.df.drop(columns=cols_to_drop, errors='ignore')
+        except Exception as e:
+            print(f"[Warning] Failed dropping columns: {e}")
+        return self
+
+    def safe_fillna(self, fill_map: dict):
+        """Fill NaN values for specific columns safely."""
+        for col, val in fill_map.items():
+            try:
+                if col in self.df.columns:
+                    self.df[col] = self.df[col].fillna(val)
+            except Exception as e:
+                print(f"[Warning] Failed filling NaN for {col}: {e}")
+        return self
+
+    def safe_replace(self, col, replacements: dict):
+        """Replace values in a column safely."""
+        try:
+            if col in self.df.columns:
+                self.df[col] = self.df[col].replace(replacements)
+        except Exception as e:
+            print(f"[Warning] Failed replacing values in {col}: {e}")
+        return self
+
+    def safe_convert_dtype(self, col, dtype, errors="ignore"):
+        """Convert column dtype safely."""
+        try:
+            if col in self.df.columns:
+                if "datetime" in str(dtype):
+                    self.df[col] = pd.to_datetime(self.df[col], errors="coerce")
+                else:
+                    self.df[col] = self.df[col].astype(dtype, errors=errors)
+        except Exception as e:
+            print(f"[Warning] Failed dtype conversion on {col}: {e}")
+        return self
+
+    def normalize_gender(self):
+        """Unify transgender categories safely."""
+        try:
+            if "Gender" in self.df.columns:
+                self.df["Gender"] = self.df["Gender"].replace({
+                    "Transgender male to female": "Transgender",
+                    "Transgender female to male": "Transgender"
+                })
+        except Exception as e:
+            print(f"[Warning] Failed gender normalization: {e}")
+        return self
+
+    def split_race(self):
+        """Split Race column into Race_1, Race_2, etc., if it exists."""
+        try:
+            if "Race" in self.df.columns:
+                splitting = self.df["Race"].astype(str).str.split(";", expand=True)
+                splitting.columns = [f"Race_{i+1}" for i in range(splitting.shape[1])]
+                self.df = pd.concat([self.df.drop(columns=["Race"]), splitting], axis=1)
+        except Exception as e:
+            print(f"[Warning] Failed race splitting: {e}")
+        return self
+
+    def clean_salary(self):
+        """Fix salary inconsistencies."""
+        try:
+            if "Salary" in self.df.columns:
+                self.df["Salary"] = pd.to_numeric(self.df["Salary"], errors="coerce")
+                self.df["Salary"] = self.df["Salary"].replace(60000, 28.84)
+        except Exception as e:
+            print(f"[Warning] Failed salary cleaning: {e}")
+        return self
+
+    def finalize(self):
+        """Return cleaned dataframe."""
+        return self.df