diff --git a/code/07_05_random_forest.py b/code/07_05_random_forest.py
index 7f5b8a3..3aa5d52 100644
--- a/code/07_05_random_forest.py
+++ b/code/07_05_random_forest.py
@@ -15,6 +15,11 @@
 cont_vars = ['dist', 'st_x', 'st_y', 'period_time_remaining', 'empty']
 cat_vars = ['pos', 'hand', 'period']
 
+# replace periods 1, 2, 3 ... with P1, P2, P3 ...
+# this is so that when we turn them into dummy variables the column names are
+# P1, P2, ... and not just 1, 2, which can cause issues
+df['period'] = 'P' + df['period'].astype(str)
+
 df_cat = pd.concat([pd.get_dummies(df[x]) for x in cat_vars], axis=1)
 
 df_all = pd.concat([df[cont_vars], df_cat], axis=1)