diff --git a/africat/aa_create_dataset.py b/africat/aa_create_dataset.py
index a93045b..9564244 100755
--- a/africat/aa_create_dataset.py
+++ b/africat/aa_create_dataset.py
@@ -105,6 +105,9 @@ def parse_and_extract(input_dir, verbose):
           text = list()
           lang = ""
           try:
+            for t in doc.find('./publisher_headline'):
+              if t.text is not None:
+                text.append(t.text)
             for p in doc.find('./body'):
               if p.text is not None:
                 text.append(p.text)
@@ -152,8 +155,8 @@ def scrub_data(articles, verbose):
   data['content'] = data.content.parallel_apply(lambda x: x.translate(remove_digits))
 
   # Remove extra spaces
-  data['content']=data.content.parallel_apply(lambda x: x.strip())
-  data['content']=data.content.parallel_apply(lambda x: re.sub(" +", " ", x))
+  data['content'] = data.content.parallel_apply(lambda x: x.strip())
+  data['content'] = data.content.parallel_apply(lambda x: re.sub(" +", " ", x))
 
   # TODO: lemmas? See spaCy
 
@@ -184,6 +187,7 @@ def main():
   articles = parse_and_extract(args.input, args.verbose)
 
   data = scrub_data(articles, args.verbose)
+  #print(data)
 
   write_csv(data, args.output)