Remove invalid responses

data-preparation.py

+import re
 from typing import Dict, List, Set
 
 import matplotlib.pyplot as plt
@@ -44,6 +45,20 @@ def _fix_typos(data_df: pd.DataFrame) -> pd.DataFrame:
     return data_df
 
 
+def _remove_answers_from_non_developers(data_df: pd.DataFrame) -> pd.DataFrame:
+    # cases in which respondents stated that they were not developing software
+    # and hence could not answer the respective questions properly
+    non_developers = [75, 100, 266, 324, 442, 506, 566, 572]
+    questions = [
+        question for question in data_df.columns if re.match(
+            r"^Q020|Q021|Q022|Q025|Q026|Q027|Q028|Q029|Q038|Q039|Q040|Q041|" +
+            "Q042|Q043|Q044", question
+        )
+    ]
+    data_df.loc[non_developers, questions] = None
+    return data_df
+
+
 def _prepare_q001(data_df: pd.DataFrame) -> pd.DataFrame:
     """Categorize affiliation mentioned in free text answers keeping "Other"
     as default category."""
@@ -1345,3 +1360,4 @@ data = _prepare_q037(data_df=data)
 data = _prepare_q038(data_df=data)
 data = _prepare_q043(data_df=data)
 data = _prepare_q044(data_df=data)
+data = _remove_answers_from_non_developers(data_df=data)