# Get the columns with > 50% missing
missing_df = missing_values_table(df_train);
missing_columns = list(missing_df[missing_df['% of Total Values'] > 50].index)
print('\n','%d columns will be deleted.' % len(missing_columns))

# Drop the columns with 50% missing data
df_train = df_train.drop(columns = list(missing_columns))