Bfill not working after changing empty strings to nan

Question

I cant get bfill to work, it continues to be nan after trying to use bfill, here is my code:

import pandas as pd


data_df = pd.read_json("Data_test.json")

pd.set_option('display.max_columns', None)
pd.set_option('display.width', None)
pd.set_option('display.max_colwidth', None)


# Replaces empty strings with nan
df2 = data_df.mask(data_df == "")

#bfill function to fill missing data with prior data. 
df2.fillna(method="bfill", inplace=True)


print(df2[["Restaurant", "Food_cat"]])

this is the results i am getting, it continue to be nan, i am trying to use bfill to fill the bottom 2 with "Most Popular" as well.

   Restaurant      Food_cat
0  McDonald's  Most Popular
1  McDonald's  Most Popular
2  McDonald's           NaN
3  McDonald's           NaN