use norm

dp_creator_ii/mock_data.py

@@ -1,5 +1,6 @@
 from collections import namedtuple
 import polars as pl
+from scipy.stats import norm
 
 ColumnDef = namedtuple("ColumnDef", ["min", "max"])
 
@@ -9,7 +10,7 @@ def mock_data(column_defs, row_count=1000):
     data = {column_name: [] for column_name in column_defs.keys()}
     for i in range(row_count):
         for column_name, column_def in column_defs.items():
-            scale = column_def.max - column_def.min
-            value = scale * i / row_count + column_def.min
+            quantile = i / row_count / 2 + 0.25  # ie, 25th to 75th percentiles
+            value = norm.ppf(quantile)
             data[column_name].append(value)
     return pl.DataFrame(data=data, schema=schema)