COLLABORATIVE FILTERING ON BOOKS¶

Installing Libraries¶

In [1]:

import pandas as pd

%matplotlib inline
import matplotlib.pyplot as plt
import numpy as np
from numpy import int64
import warnings
warnings.filterwarnings("ignore")

import requests
import IPython.display as Disp
import plotly.express as px

import sklearn
from IPython.core.display import display,HTML
from sklearn.decomposition import TruncatedSVD

Data Reading¶

In [2]:

# Books data reading
book=pd.read_csv("C:Books.txt") 

In [3]:

#Rating data reading 
rating=pd.read_csv("C:Ratings.txt")

In [4]:

book.shape,rating.shape

Out[4]:

((10000, 23), (162604, 3))

In [5]:

pd.set_option('display.max_columns',100)
book.head(3)

Out[5]:

	book_id	goodreads_book_id	best_book_id	work_id	books_count	isbn	isbn13	authors	original_publication_year	original_title	title	language_code	average_rating	ratings_count	work_ratings_count	work_text_reviews_count	ratings_1	ratings_2	ratings_3	ratings_4	ratings_5	image_url	small_image_url
0	1	2767052	2767052	2792775	272	439023483	9.780439e+12	Suzanne Collins	2008.0	The Hunger Games	The Hunger Games (The Hunger Games, #1)	eng	4.34	4780653	4942365	155254	66715	127936	560092	1481305	2706317	https://images.gr-assets.com/books/1447303603m...	https://images.gr-assets.com/books/1447303603s...
1	2	3	3	4640799	491	439554934	9.780440e+12	J.K. Rowling, Mary GrandPré	1997.0	Harry Potter and the Philosopher's Stone	Harry Potter and the Sorcerer's Stone (Harry P...	eng	4.44	4602479	4800065	75867	75504	101676	455024	1156318	3011543	https://images.gr-assets.com/books/1474154022m...	https://images.gr-assets.com/books/1474154022s...
2	3	41865	41865	3212258	226	316015849	9.780316e+12	Stephenie Meyer	2005.0	Twilight	Twilight (Twilight, #1)	en-US	3.57	3866839	3916824	95009	456191	436802	793319	875073	1355439	https://images.gr-assets.com/books/1361039443m...	https://images.gr-assets.com/books/1361039443s...

In [6]:

rating.head(3)

Out[6]:

	user_id	book_id	rating
0	1	258	5
1	2	4081	4
2	2	260	5

Data Merging¶

In [7]:

# Data merging done on common column i.e. book id
df=rating.merge(book,on='book_id')

In [8]:

df.head(2)

Out[8]:

	user_id	book_id	rating	goodreads_book_id	best_book_id	work_id	books_count	isbn	isbn13	authors	original_publication_year	original_title	title	language_code	average_rating	ratings_count	work_ratings_count	work_text_reviews_count	ratings_1	ratings_2	ratings_3	ratings_4	ratings_5	image_url	small_image_url
0	1	258	5	1232	1232	3209783	279	143034901	9.780143e+12	Carlos Ruiz Zafón, Lucia Graves	2001.0	La sombra del viento	The Shadow of the Wind (The Cemetery of Forgot...	eng	4.24	263685	317554	24652	4789	11769	42214	101612	157170	https://images.gr-assets.com/books/1344545047m...	https://images.gr-assets.com/books/1344545047s...
1	11	258	3	1232	1232	3209783	279	143034901	9.780143e+12	Carlos Ruiz Zafón, Lucia Graves	2001.0	La sombra del viento	The Shadow of the Wind (The Cemetery of Forgot...	eng	4.24	263685	317554	24652	4789	11769	42214	101612	157170	https://images.gr-assets.com/books/1344545047m...	https://images.gr-assets.com/books/1344545047s...

In [9]:

df.describe()

Out[9]:

	user_id	book_id	rating	goodreads_book_id	best_book_id	work_id	books_count	isbn13	original_publication_year	average_rating	ratings_count	work_ratings_count	work_text_reviews_count	ratings_1	ratings_2	ratings_3	ratings_4	ratings_5
count	162604.000000	162604.000000	162604.000000	1.626040e+05	1.626040e+05	1.626040e+05	162604.000000	1.607660e+05	162579.000000	162604.000000	1.626040e+05	1.626040e+05	162604.000000	162604.00000	162604.000000	162604.000000	1.626040e+05	1.626040e+05
mean	2085.741765	1772.377119	3.813184	1.333508e+05	4.646940e+05	2.359180e+06	305.715032	9.776478e+12	1945.111823	3.974965	4.176416e+05	4.464290e+05	10714.792059	11066.13202	22563.078842	77512.512386	1.404566e+05	1.948307e+05
std	1212.894799	2361.275403	1.001269	9.495106e+05	2.865794e+06	4.224005e+06	503.339815	1.988365e+11	220.054157	0.254314	6.719332e+05	7.053673e+05	14638.133223	22737.33095	37251.600050	112010.698268	2.066473e+05	3.699612e+05
min	1.000000	2.000000	1.000000	1.000000e+00	1.000000e+00	8.700000e+01	1.000000	1.951703e+08	-1750.000000	2.470000	2.716000e+03	5.510000e+03	17.000000	11.00000	30.000000	323.000000	7.500000e+02	7.810000e+02
25%	1037.000000	136.000000	3.000000	5.211000e+03	5.348000e+03	9.007640e+05	60.000000	9.780307e+12	1955.000000	3.820000	4.006300e+04	4.421500e+04	1587.000000	682.00000	2227.000000	9745.000000	1.580200e+04	1.414800e+04
50%	2048.000000	680.000000	4.000000	1.493500e+04	1.588100e+04	1.758256e+06	114.000000	9.780440e+12	1989.000000	3.970000	1.310040e+05	1.423260e+05	4587.000000	2631.00000	7716.000000	29992.000000	4.949300e+04	4.757200e+04
75%	3101.000000	2465.250000	5.000000	5.203600e+04	6.496000e+04	2.977639e+06	283.000000	9.780685e+12	2000.000000	4.140000	4.698360e+05	5.105820e+05	13050.000000	10381.00000	24094.000000	93427.000000	1.654770e+05	1.875690e+05
max	5518.000000	10000.000000	5.000000	1.974776e+07	3.360215e+07	5.639960e+07	3455.000000	9.790008e+12	2014.000000	4.820000	4.602479e+06	4.800065e+06	95009.000000	456191.00000	436802.000000	793319.000000	1.156318e+06	3.011543e+06

In [10]:

df.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 162604 entries, 0 to 162603
Data columns (total 25 columns):
 #   Column                     Non-Null Count   Dtype  
---  ------                     --------------   -----  
 0   user_id                    162604 non-null  int64  
 1   book_id                    162604 non-null  int64  
 2   rating                     162604 non-null  int64  
 3   goodreads_book_id          162604 non-null  int64  
 4   best_book_id               162604 non-null  int64  
 5   work_id                    162604 non-null  int64  
 6   books_count                162604 non-null  int64  
 7   isbn                       160413 non-null  object 
 8   isbn13                     160766 non-null  float64
 9   authors                    162604 non-null  object 
 10  original_publication_year  162579 non-null  float64
 11  original_title             160786 non-null  object 
 12  title                      162604 non-null  object 
 13  language_code              151898 non-null  object 
 14  average_rating             162604 non-null  float64
 15  ratings_count              162604 non-null  int64  
 16  work_ratings_count         162604 non-null  int64  
 17  work_text_reviews_count    162604 non-null  int64  
 18  ratings_1                  162604 non-null  int64  
 19  ratings_2                  162604 non-null  int64  
 20  ratings_3                  162604 non-null  int64  
 21  ratings_4                  162604 non-null  int64  
 22  ratings_5                  162604 non-null  int64  
 23  image_url                  162604 non-null  object 
 24  small_image_url            162604 non-null  object 
dtypes: float64(3), int64(15), object(7)
memory usage: 32.3+ MB

Data Cleaning¶

In [11]:

df.isna().sum()

Out[11]:

user_id                          0
book_id                          0
rating                           0
goodreads_book_id                0
best_book_id                     0
work_id                          0
books_count                      0
isbn                          2191
isbn13                        1838
authors                          0
original_publication_year       25
original_title                1818
title                            0
language_code                10706
average_rating                   0
ratings_count                    0
work_ratings_count               0
work_text_reviews_count          0
ratings_1                        0
ratings_2                        0
ratings_3                        0
ratings_4                        0
ratings_5                        0
image_url                        0
small_image_url                  0
dtype: int64

In [12]:

#Filling null values of original_title with Na and 0
df['original_title']=df['original_title'].fillna('Na') 
df['original_publication_year']=df['original_publication_year'].fillna(0)

In [13]:

#Since, all columns of data are not required. So, extracting the columns that are only necessary
dfuse=df[['user_id','book_id','rating','books_count','original_publication_year','average_rating','ratings_count','title','original_title','authors','image_url']]

In [14]:

dfuse.head(3)

Out[14]:

	user_id	book_id	rating	books_count	original_publication_year	average_rating	ratings_count	title	original_title	authors	image_url
0	1	258	5	279	2001.0	4.24	263685	The Shadow of the Wind (The Cemetery of Forgot...	La sombra del viento	Carlos Ruiz Zafón, Lucia Graves	https://images.gr-assets.com/books/1344545047m...
1	11	258	3	279	2001.0	4.24	263685	The Shadow of the Wind (The Cemetery of Forgot...	La sombra del viento	Carlos Ruiz Zafón, Lucia Graves	https://images.gr-assets.com/books/1344545047m...
2	143	258	4	279	2001.0	4.24	263685	The Shadow of the Wind (The Cemetery of Forgot...	La sombra del viento	Carlos Ruiz Zafón, Lucia Graves	https://images.gr-assets.com/books/1344545047m...

Visualisation¶

Q1 Maximum Count of Ratings are of ?¶

In [15]:

px.bar(dfuse.groupby('rating')['book_id'].count(), title='Plot of Rating with Books Count',color_discrete_sequence =['green']*3, height=500, width=700)

As shown here, maximum rating of 4 are given 58.821K times where rating 5 are given 46.31K. Least rating given is 1 to only 3633. This shows given in data are good books since rating 3 are 4 times given than rating 2.

Q2 Top 5 Books which are Rated in High Count ?¶

In [16]:

px.bar(dfuse.groupby('book_id')['rating'].count().sort_values(ascending=False).nlargest(5), title='Plot of Top 5 Highly Rated in Counts',color_discrete_sequence =['light blue']*3, height=500, width=700 )

As shown maximum rating count is of book id 26 with 1010 count followed by count of 922 of book id of 2 and 4.

Q3 Which are Top Highly Rated Books Based on Average Rating ?¶

In [17]:

px.bar(dfuse.groupby(str('title'))['average_rating'].mean().sort_values(ascending=False).nlargest(5), title='Plot of Top 3 Highly Rated Books on Basis of Average Ratings',color_discrete_sequence =['red']*3,height=500, width=700 )

As shown above, highest rating average of 4.82 is given to book "The Complete Calvin and Hobbies" followed by rating of 4.77 to Harry Potter(1-5)

Q4 Which User Rated Maximum ?¶

In [18]:

px.bar(dfuse.groupby(str('user_id'))['rating'].count().sort_values(ascending=False).nlargest(10), title='Plot of User Id with Ratings Count',color_discrete_sequence =['orange']*3 , height=500, width=700)

As shown above, user id of 2276 rated maximum books of 185 in count followed by value of 179 by user if 4147 then value of 172 with user id 1794

Q5 Which Original Title has Maximum Rating ?¶

In [19]:

df_1=pd.DataFrame(dfuse.groupby('original_title')['rating'].count().sort_values(ascending=False).nlargest(6))

In [20]:

df_1['Image']=" "
def path_to_image_html(path):
    return '<img src="'+ path + '" width="60" >'

In [21]:

for i in range(1,len(df_1.index)):
    url=dfuse[dfuse['original_title']==df_1.index[i]]['image_url'].unique()
    df_1['Image'][i]=url[0]

In [22]:

image_cols = ['Image']

# Create the dictionariy to be passed as formatters
format_dict = {}
for image_col in image_cols:
    format_dict[image_col] = path_to_image_html

display(HTML(df_1[1:].to_html(escape=False ,formatters=format_dict)))

	rating	Image
original_title
The Da Vinci Code	1010
To Kill a Mockingbird	922
Harry Potter and the Philosopher's Stone	922
The Catcher in the Rye	883
Harry Potter and the Prisoner of Azkaban	883

As given in above dataframe, "The Da Vinci Code" has the maximum rating count of 1010 haing book id 26 as mentioned in Q2

Recommendation System¶

In [23]:

# create pivot table to study the user and book relation
df_pivot=dfuse.pivot_table(values='rating',index='user_id',columns='original_title',fill_value=0)

In [24]:

df_pivot.head(5)

Out[24]:

original_title		Animal Farm & 1984	Burned (Burned, #1)	Carter Beats the Devil	De Imitatione Christi	Faust. Der Tragödie erster und zweiter Teil	Feeling Good: The New Mood Therapy	Innocent Traitor	The Fellowship of the Ring	The Sea of Trolls	The Tale of Despereaux	Игрок	الرحيق المختوم: بحث في السيرة النبوية على صاحبها أفضل الصلاة والسلام	سمفونی مردگان	めくらやなぎと、眠る女 [Mekurayanagi to, nemuru onna]	国境の南、太陽の西 [Kokkyō no minami, taiyō no nishi]	絶対彼氏 (Zettai Kareshi) 1	"...And Ladies of the Club"	"D" is for Deadbeat	"M" is for Malice	"N" is for Noose	"O" is for Outlaw	"P" is for Peril	"Q" is for Quarry	'Tis: A Memoir	... Then He Ate My Boy Entracers	...and that's when it fell off in my hand	1,000 Places to See Before You Die	100 Bullets, Vol. 1: First Shot, Last Call (100 bullets, #1)	100 Selected Poems by e. e. Cummings	13 Little Blue Envelopes	1421: The Year China Discovered America	1491: New Revelations of the Americas Before Columbus	16 Lighthouse Road	1776	1st To Die	2001: A Space Odyssey	2010: Odyssey Two	204 Rosewood Lane	2061: Odyssey Three	20th Century Ghosts	2666	2nd Chance	30 Days of Night	300	3001: The Final Odyssey	31 Songs	3rd Degree	4.50 From Paddington	44 Scotland Street	...	デスノート #2 (Desu Nōto) Gōryū (合流)	デスノート #3 (Desu Nōto) Gekisō (激走)	デスノート #4 (Desu Nōto) Koigokoro (恋心)	デスノート #5 (Desu Nōto) Hakushi (白紙,)	デスノート #6 (Desu Nōto) Kōkan (交換)	デスノート #7 (Desu Nōto) Zero (零)	デスノート #8 (Desu Nōto) Mato (的)	ノルウェイの森 [Noruwei no Mori]	バトル・ロワイアル	フルーツバスケット 1	フルーツバスケット 14	フルーツバスケット 15	フルーツバスケット 2	フルーツバスケット 3	フルーツバスケット 4	フルーツバスケット 7	マース (Mars) #1	ヤマトナデシコ七変化	ラブひな #1	ヴァンパイア騎士 1	ヴァンパイア騎士 2	ヴァンパイア騎士 3	ヴァンパイア騎士 4	世界の終りとハードボイルド・ワンダーランド [Sekai no owari to hādoboirudo wandārando]	五輪書 [Go Rin no Sho]	午後の曳航 [Gogo no eikō]	名探偵コナン 1	孫子兵法 [Sūnzi bīngfǎ]	宮本武蔵 Miyamoto Musashi	易 [Yì]	桜蘭高校ホスト部 1	桜蘭高校ホスト部 2	桜蘭高校ホスト部 3	桜蘭高校ホスト部 4	沈黙	海辺のカフカ [Umibe no Kafuka]	満月をさがして 1	犬夜叉 1	砂の女 [Suna no onna]	神の子どもたちはみな踊る	窓ぎわのトットちゃん	紳士同盟† 1	羊をめぐる冒険 [Hitsujiwomegurubōken]	美少女戦士セーラームーン 1 [Bishōjo Senshi Sailor Moon 1]	花より男子 1	論語 [Lún Yǔ]	象の消滅 [Zō no shōmetsu]	道德經 [dào dé jīng]	鋼の錬金術師 1	雪国
user_id
1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
2	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
3	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
4	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
6	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	...	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0

5 rows × 5286 columns

In [25]:

# Transposing the pivot table to easy access
x=df_pivot.values.T  
x.shape 

Out[25]:

(5286, 2780)

Dimentionality Reduction Using SVD¶

In [26]:

SVD  = TruncatedSVD(n_components=20, random_state=17)
result_matrix = SVD.fit_transform(x)
result_matrix.shape

Out[26]:

(5286, 20)

Creating the Pearson's Correlation¶

In [27]:

corr_mat = np.corrcoef(result_matrix)
corr_mat.shape

Out[27]:

(5286, 5286)

Recommendation of The Entered Book¶

In [28]:

book_names = df_pivot.columns
book_list = list(book_names)
book_index = book_list.index(' Innocent Traitor') 

In [29]:

corr_book = corr_mat[book_index] 
corr_book.shape

Out[29]:

(5286,)

In [30]:

print("Recommendation are")
list(book_names[(corr_book<1.0) & (corr_book>0.8)][1:])

Recommendation are

Out[30]:

['A Tree Grows In Brooklyn ',
 'Breathing Lessons',
 'Drowning Ruth',
 "French Women Don't Get Fat",
 'Girl with a Pearl Earring',
 'September',
 'The Accidental Tourist',
 "The Bonesetter's Daughter",
 'The Joy Luck Club',
 'The Last Days of Dogtown',
 'The Prince of Tides',
 'We Were the Mulvaneys',
 'Winter Solstice']

Function for Recommendation System Gives Top 5 Books¶

In [31]:

def path_to_image_html(path):
    return '<img src="'+ path + '" width="60" >'

def df_recommend(recommend):
    recommend=recommend[:5]
    year=[]
    image_url=[]
    for i in recommend:
        for j in dfuse.index:
            if dfuse['original_title'][j]==i:
                year.append(dfuse['original_publication_year'][j])
                image_url.append(dfuse['image_url'][j])  
                break
    recommend_df=pd.DataFrame([recommend,year,image_url]).T
    recommend_df.columns=['Recommend Books','Year of Publication','Image']
    
    image_cols = ['Image']


    format_dict = {}
    for image_col in image_cols:
        format_dict[image_col] = path_to_image_html

    display(HTML(recommend_df[0:10].to_html(escape=False ,formatters=format_dict)))

In [32]:

def recommend_book(df_pivot, corr_mat):
    name=input("Enter the Name of the Book : ")
    book_names = df_pivot.columns
    book_list = list(book_names)
    try:
        if name in book_list:
            book_index = book_list.index(name) 
            corr_book = corr_mat[book_index] 
            print("Recommendation are")
            recommend=list(book_names[(corr_book<1.0) & (corr_book>0.8)][1:])
            df_recommend(recommend)

        else:
            name=" "+name
            book_index = book_list.index(name) 
            corr_book = corr_mat[book_index] 
            print("Recommendation are")
            recommend=list(book_names[(corr_book<1.0) & (corr_book>0.8)][1:])
            df_recommend(recommend)
    except:
        print("Enter the Book Name Again")
        recommend_book(df_pivot,corr_mat)    
    

In [34]:

recommend_book(df_pivot,corr_mat)   

Enter the Name of the Book : The Prince of Tides
Recommendation are

	Recommend Books	Year of Publication
0	A Tree Grows In Brooklyn	1943
1	All Over But the Shoutin'	1997
2	Angela's Ashes: A Memoir	1996
3	Back When We Were Grownups	2001
4	Beach Music	1975

In [ ]: