How to calculate the mean of ratings of each user?

问题

Assume I have a dataset like this:

userID productID rating
a          i       5
b          i       4
c          i       4 
a          j       3
b          j       5

The question is, how can I calculate the mean rating of each user? I saw this answer, but I didn't quite understand it. I would really appreciate your time, if you show some guidance.

回答1:

I work in an IPython Notebook.

Let's assume you have this file user_ratings.csv:

userID productID rating
a          i       5
b          i       4
c          i       4
a          j       3
b          j       5

The example in the link uses pandas. So import pandas:

In [1]: import pandas as pd

Read your file into a dataframe:

In [2]: df = pd.read_csv('user_ratings.csv', delim_whitespace=True)
df 

Group by the user and calculate the mean for each:

In [2]: df.groupby('userID').mean()

You can also create a new column in df named user_avg_rating an assign the mean score of each user to it:

In [3]:  df['user_avg_rating'] = df.groupby('userID')['rating'].transform('mean')
         df

The method transform takes your grouped object and creates a series:

In [4]:  df.groupby('userID')['rating'].transform('mean')

    0    4.0
    1    4.5
    2    4.0
    3    4.0
    4    4.5
    dtype: float64

This series is assigned to the column user_avg_rating.

来源：https://stackoverflow.com/questions/34409600/how-to-calculate-the-mean-of-ratings-of-each-user