diff --git a/src/main/python/calibrate.py b/src/main/python/calibrate.py
index 3ffcdf0..cb583c5 100644
--- a/src/main/python/calibrate.py
+++ b/src/main/python/calibrate.py
@@ -33,7 +33,7 @@
     "ride": -0.76
 }
 
-# Based on MiD 2017, filtered on Lausitz region
+# Based on MiD 2017, filtered on Lausitz region (see create_ref.py)
 target = {
     "walk": 0.199819,
     "bike": 0.116362,
diff --git a/src/main/python/create_ref.py b/src/main/python/create_ref.py
index 5f81576..9bd9c27 100644
--- a/src/main/python/create_ref.py
+++ b/src/main/python/create_ref.py
@@ -11,27 +11,36 @@
 
 
 def person_filter(df):
+    """ Filter person that are relevant for the calibration."""
+
     df = gpd.GeoDataFrame(df, geometry=gpd.GeoSeries.from_wkt(df.geom, crs="EPSG:4326").to_crs(CRS))
     df = gpd.sjoin(df, region, how="inner", predicate="intersects")
 
+    # Groups will be shown on the dashboard
     df["age"] = cut(df.age, [0, 12, 18, 25, 35, 66, np.inf])
+    df["hh_income"] = cut(df.income, [0, 500, 900, 1500, 2000, 2600, 3000, 3600, 4600, 5600, np.inf])
 
+    # Only weekdays are considered, with persons present in their home region
     return df[df.present_on_day & (df.reporting_day <= 5)]
 
 
 def trip_filter(df):
-    # Other modes are ignored in the total share
+    # All modes, expect for "other" are considered
     return df[df.main_mode != "other"]
 
 
 if __name__ == "__main__":
+
+    # Defines the Lausitz region
     region = gpd.read_file("../../../../shared-svn/projects/DiTriMo/data/shp/lausitz.shp").to_crs(CRS)
 
+    # This contains the path to the MiD 2017 data with the highest resolution
+    # See https://daten.clearingstelle-verkehr.de/279/ for more information, the data is not included in this repository
     r = run_create_ref_data.create(
         "/Volumes/Untitled/B3_Lokal-Datensatzpaket/CSV",
         person_filter, trip_filter,
         run_create_ref_data.InvalidHandling.REMOVE_TRIPS,
-        ref_groups=["age"]
+        ref_groups=["age", "hh_income", "economic_status"]
     )
 
     print("Filtered %s persons" % len(r.persons))
@@ -39,7 +48,9 @@ def trip_filter(df):
 
     print(r.share)
 
-    # Simulated trips
+    # Calculate the number of short distance trips that are missing in the simulated data
+    # This function required that one run with 0 iterations has been performed beforehand
+
     sim_persons = pd.read_csv("../../../output/output-lausitz-100pct/lausitz-100pct.output_persons.csv.gz",
                               delimiter=";", dtype={"person": "str"})
     sim_persons = sim_persons[sim_persons.subpopulation == "person"]
@@ -53,5 +64,5 @@ def trip_filter(df):
 
     sim = pd.merge(sim, sim_persons, how="inner", left_on="person", right_on="person", validate="many_to_one")
 
-    share, add_trips = calc_needed_short_distance_trips(trips, sim, max_dist=700)
+    share, add_trips = calc_needed_short_distance_trips(r.trips, sim, max_dist=700)
     print("Short distance trip missing: ", add_trips)