diff --git a/src/python/main/evaluation_results.py b/src/python/main/evaluation_results.py
index e91c51e..cb6b79a 100644
--- a/src/python/main/evaluation_results.py
+++ b/src/python/main/evaluation_results.py
@@ -6,12 +6,17 @@
 from typing import List
 
 import pandas as pd
-
-from . import tangled_metrics
-
 from pandas.api.types import CategoricalDtype
 
 
+TANGLED_LEVELS = [
+    "tangled_lines",
+    "tangled_hunks",
+    "tangled_files",
+    "tangled_patch",
+    "single_concern_patch",
+]
+
 GROUND_TRUTH_COLUMNS = ["file", "source", "target", "group"]
 PERFORMANCE_COLUMNS = [
     "project",
@@ -62,7 +67,7 @@ def read_metrics(file: str, dataset_name=None) -> pd.DataFrame:
 
     if "tangled_level" in df.columns:
         df["tangled_level"] = df["tangled_level"].astype(
-            CategoricalDtype(categories=tangled_metrics.TANGLED_LEVELS, ordered=True)
+            CategoricalDtype(categories=TANGLED_LEVELS, ordered=True)
         )
 
     # Convert the commit_id column to string for D4J bug ids.
diff --git a/src/python/main/tangled_metrics.py b/src/python/main/tangled_metrics.py
index 00b9151..1612896 100644
--- a/src/python/main/tangled_metrics.py
+++ b/src/python/main/tangled_metrics.py
@@ -10,18 +10,10 @@
 import argparse
 import pandas as pd
 import numpy as np
+from pandas.api.types import CategoricalDtype
 
 from . import metrics
 from . import evaluation_results
-from pandas.api.types import CategoricalDtype
-
-TANGLED_LEVELS = [
-    "tangled_lines",
-    "tangled_hunks",
-    "tangled_files",
-    "tangled_patch",
-    "single_concern_patch",
-]
 
 
 def main(metrics_file: str, results_dir: str):
@@ -98,7 +90,7 @@ def calculate_tangled_levels(dataframe: pd.DataFrame) -> pd.DataFrame:
 
     dataframe["tangled_level"] = None
     # Go through each level from finer granularity to coarser granularity and set the tangled level for each commit if it is not set yet.
-    for level in TANGLED_LEVELS:
+    for level in evaluation_results.TANGLED_LEVELS:
         dataframe["tangled_level"] = np.where(
             (dataframe[level] > 0) & (dataframe["tangled_level"].isnull()),
             level,
@@ -106,7 +98,7 @@ def calculate_tangled_levels(dataframe: pd.DataFrame) -> pd.DataFrame:
         )
 
     dataframe["tangled_level"] = dataframe["tangled_level"].astype(
-        CategoricalDtype(categories=TANGLED_LEVELS, ordered=True)
+        CategoricalDtype(categories=evaluation_results.TANGLED_LEVELS, ordered=True)
     )
 
     return dataframe