Python for Recommender Systems
Python code in R Markdown
Paul Jozefek
2020-09-28
Recommender Systems
Objective
- Develop basic recommendation systems using Python and pandas.
- Suggesting items that are most similar to a particular item. In this case, movies.
This is not a true robust recommendation system. It generates movies/items that are most similar to your choice.
> import numpy as np
+ import pandas as pd
+ import matplotlib.pyplot as plt
+ import seaborn as snsThe Data
> column_names = ['user_id', 'item_id', 'rating', 'timestamp']
> df = pd.read_csv('u.data', sep='\t', names=column_names) user_id item_id rating timestamp
0 0 50 5 881250949
1 0 172 5 881250949
2 0 133 1 881250949
3 196 242 3 881250949
4 186 302 3 891717742Now let’s get the movie titles:
item_id title
0 1 Toy Story (1995)
1 2 GoldenEye (1995)
2 3 Four Rooms (1995)
3 4 Get Shorty (1995)
4 5 Copycat (1995)We can merge them together:
user_id item_id rating timestamp title
0 0 50 5 881250949 Star Wars (1977)
1 290 50 5 880473582 Star Wars (1977)
2 79 50 4 891271545 Star Wars (1977)
3 2 50 5 888552084 Star Wars (1977)
4 8 50 5 879362124 Star Wars (1977)Exploratory Data Analysis
Let’s create a ratings dataframe with average rating and number of ratings:
title
Marlene Dietrich: Shadow and Light (1996) 5.0
Prefontaine (1997) 5.0
Santa with Muscles (1996) 5.0
Star Kid (1997) 5.0
Someone Else's America (1995) 5.0
Name: rating, dtype: float64title
Star Wars (1977) 584
Contact (1997) 509
Fargo (1996) 508
Return of the Jedi (1983) 507
Liar Liar (1997) 485
Name: rating, dtype: int64 rating
title
'Til There Was You (1997) 2.333333
1-900 (1994) 2.600000
101 Dalmatians (1996) 2.908257
12 Angry Men (1957) 4.344000
187 (1997) 3.024390Now set the number of ratings column:
> ratings['num of ratings'] = pd.DataFrame(
+ df.groupby('title')['rating'].count())
+ ratings.head() rating num of ratings
title
'Til There Was You (1997) 2.333333 9
1-900 (1994) 2.600000 5
101 Dalmatians (1996) 2.908257 109
12 Angry Men (1957) 4.344000 125
187 (1997) 3.024390 41Now a few histograms:
> sns.set_style('whitegrid')
+ plt.figure(figsize=(10,4))
+ ratings['num of ratings'].hist(bins=70);
+ plt.show()
> sns.jointplot(x='rating',y='num of ratings',
+ data=ratings,alpha=0.5);
+ plt.tight_layout()
+ plt.show()
Recommending Similar Movies
Now let’s create a matrix that has the user ids on one access and the movie title on another axis. Each cell will then consist of the rating the user gave to that movie. Note there will be a lot of NaN values, because most people have not seen most of the movies.
> moviemat = df.pivot_table(index='user_id',
+ columns='title',values='rating')
+ moviemat.iloc[0:5,0:4]
+ # display first 4 columnstitle 'Til There Was You (1997) ... 12 Angry Men (1957)
user_id ...
0 NaN ... NaN
1 NaN ... 5.0
2 NaN ... NaN
3 NaN ... NaN
4 NaN ... NaN
[5 rows x 4 columns]1664Most rated movie:
rating num of ratings
title
Star Wars (1977) 4.359589 584
Contact (1997) 3.803536 509
Fargo (1996) 4.155512 508
Return of the Jedi (1983) 4.007890 507
Liar Liar (1997) 3.156701 485
English Patient, The (1996) 3.656965 481
Scream (1996) 3.441423 478
Toy Story (1995) 3.878319 452
Air Force One (1997) 3.631090 431
Independence Day (ID4) (1996) 3.438228 429Let’s choose two movies: starwars, a sci-fi movie. And Liar Liar, a comedy.
rating num of ratings
title
'Til There Was You (1997) 2.333333 9
1-900 (1994) 2.600000 5
101 Dalmatians (1996) 2.908257 109
12 Angry Men (1957) 4.344000 125
187 (1997) 3.024390 41Now let’s grab the user ratings for those two movies:
> starwars_user_ratings = moviemat['Star Wars (1977)']
+ liarliar_user_ratings = moviemat['Liar Liar (1997)']
+ starwars_user_ratings.head()user_id
0 5.0
1 5.0
2 5.0
3 NaN
4 5.0
Name: Star Wars (1977), dtype: float64We can then use corrwith() method to get correlations between two pandas series:
Let’s clean this by removing NaN values and using a DataFrame instead of a series:
> corr_starwars = pd.DataFrame(
+ similar_to_starwars,columns=['Correlation'])
+ corr_starwars.dropna(inplace=True)
+ corr_starwars.head() Correlation
title
'Til There Was You (1997) 0.872872
1-900 (1994) -0.645497
101 Dalmatians (1996) 0.211132
12 Angry Men (1957) 0.184289
187 (1997) 0.027398Now if we sort the dataframe by correlation, we should get the most similar movies, however note that we get some results that don’t really make sense. This is because there are a lot of movies only watched once by users who also watched star wars (it was the most popular movie).
Correlation
title
Commandments (1997) 1.0
Cosi (1996) 1.0
No Escape (1994) 1.0
Stripes (1981) 1.0
Man of the Year (1995) 1.0
Hollow Reed (1996) 1.0
Beans of Egypt, Maine, The (1994) 1.0
Good Man in Africa, A (1994) 1.0
Old Lady Who Walked in the Sea, The (Vieille qu... 1.0
Outlaw, The (1943) 1.0Let’s fix this by filtering out movies that have less than 100 reviews (this value was chosen based off the histogram from earlier).
Correlation num of ratings
title
'Til There Was You (1997) 0.872872 9
1-900 (1994) -0.645497 5
101 Dalmatians (1996) 0.211132 109
12 Angry Men (1957) 0.184289 125
187 (1997) 0.027398 41Now sort the values and notice how the titles make a lot more sense:
> corr_starwars[corr_starwars['num of ratings']>100].sort_values(
+ 'Correlation',ascending=False).round(2).head() Correlation num of ratings
title
Star Wars (1977) 1.00 584
Empire Strikes Back, The (1980) 0.75 368
Return of the Jedi (1983) 0.67 507
Raiders of the Lost Ark (1981) 0.54 420
Austin Powers: International Man of Mystery (1997) 0.38 130Now the same for the comedy Liar Liar:
> corr_liarliar = pd.DataFrame(similar_to_liarliar,
+ columns=['Correlation'])
+ corr_liarliar.dropna(inplace=True)
+ corr_liarliar = corr_liarliar.join(ratings['num of ratings'])
+ corr_liarliar[corr_liarliar['num of ratings']>100].sort_values(
+ 'Correlation',ascending=False).round(2).head() Correlation num of ratings
title
Liar Liar (1997) 1.00 485
Batman Forever (1995) 0.52 114
Mask, The (1994) 0.48 129
Down Periscope (1996) 0.47 101
Con Air (1997) 0.47 137