Data Science Project on Amazon Best Selling Books Analysis with the Python Programming Language¶
About the dataset: This dataset contains information about top selling books of Amazon(550 books) either belonging to fiction or non-fiction.
The data I'll be using in this data science project is a dataset of Amazon's 50 Best Books between 2009 and 2019.
Task 1: Reading and Inspection¶
- ### Subtask 1.1: Import and Read
Import and Read the Amazon Best Selling Book dataset and Store it in a variable called data.
Importing Libraries¶
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
Loading the Dataset¶
data = pd.read_csv("bestsellers with categories.csv")
data.head()
Subtask 1.2: Inspect the dataframe¶
Inspect the dataframe's columns, shapes, variable types, unique values etc.
Check Shape¶
r,c = data.shape
print(f"The dataset has {r} rows and {c} columns.")
Check Datatype of Each Column¶
data.dtypes
Detail Information about Dataset¶
data.info()
Check The Total Unique Values In Each Column¶
data.nunique()
Get An Overview Of The Dataframe¶
data.describe()
Note: It is clear from the above table that maximum price of a book is 105$ and maximum rating is 4.9.
Task 2: Cleaning the Data¶
- ### Subtask 2.1: Inspect Null values
Find out the number of Null values in all the columns and rows
Column-Wise Null Count¶
data.isnull().sum()
Row-Wise Null Count¶
data.isnull().sum(axis=1).sort_values(ascending=False)
Subtask 2.2: Check if the dataset has some duplicate values?¶
data[data.duplicated()]
len(data[data.duplicated()])
Note: There is no duplicate row found in the dataset
Check Total number of different books¶
len(data.Name.unique())
Note: Total number of unique books are 351 but total rows are 550.
This is clear if we see the dataframe that some books are repeatedly among the top sellers in different years.
Check The Duplicate Books¶
data[['Name','Author']].duplicated().sum()
Task 3: Data Analysis¶
### Subtask 3.1: Find the Book with highest profit
- Create a new column called
Profitwhich contains the product of the two columns:ReviewsandPrice. - Sort the dataframe using the
profitcolumn as reference. - Extract the top ten profiting Books in descending order and store them in a new dataframe -
top10
- Create a new column called
data['Profit'] = data['Reviews'] * data['Price']
data
Top10 = data.sort_values(by='Profit',ascending=False).head(10)
Top10
- ### Subtask 3.2: Drop duplicate values
After you found out the top 10 profiting books, you might have notice a duplicate value. So, it seems like the dataframe has duplicate values as well. Group the dataframe by name and repeat Subtask 3.1.
Top10 = data.groupby('Name')[['Profit']].max().sort_values(by='Profit',ascending=False).head(10)
Top10
sns.barplot(x=Top10.Profit,y=Top10.index.values)
plt.xlabel("Earned(millions)")
plt.ylabel("Books")
plt.title('Earning by Books')
- ### Subtask 3.3: Which Genre Has The Most Books In This Category and Their Distribution?
data.Genre.unique()
data.Genre.value_counts()
sns.countplot(x="Genre", data=data)
plt.show()
plt.pie(data.Genre.value_counts(),labels=['Non Fiction','Fiction'],autopct='%.0f%%');
Non Fiction books are in majority in top selling category.
- ### Subtask 3.4: What is The Average Rating of Each Genre and Plot Histogtam of each Rating
data.groupby('Genre')['User Rating'].mean()
sns.histplot(data=data['User Rating'],bins=10)
plt.xlabel("Ratings")
- ### Subtask 3.5: Find the Relationship of Ratings with Time.
sns.lineplot(y=data['User Rating'],x=data['Year'],hue=data['Genre']);
plt.ylabel("Ratings")
plt.xlabel("Years");
It is clear from the above graph that most of the books received ratings between 4.5 to 4.9.
- ### Subtask 3.6: Find Relationship Between Ratings and Price
Plot Data and Regression Model Fits Across a FacetGrid.¶
sns.lmplot(y='User Rating',x='Price',data=data)
plt.ylabel('Ratings')
plt.xlabel('Price');
Note: As we seen price increases, rating goes down.
data.groupby('Genre')['User Rating'].mean()
This graph shows that there is not a significant relationship between price and ratings but with increasing price , the ratings are falling for both Fiction and Non fiction.
- ### Subtask 3.6: Find How Much Books Have Earned Yearly.
data.groupby('Year').Profit.sum().sort_values(ascending=False)
data.groupby('Year').Profit.sum().plot(kind='bar')
plt.show()
data.groupby(['Year','Genre']).Profit.sum().sort_values(ascending=False)
Plot using Seaborn¶
sns.barplot(x=data['Year'],y=data['Profit'],hue=data['Genre'])
plt.xlabel('Years')
plt.ylabel("Eearned(millions)")
plt.title('Money earned each year');
Above graph above shows the earning of these top books ,in most cases, are below 0.8 million.
- ### Subtask 3.7: Average Profit Earned by Each Book Depending on its Genre.
genre_average=data.groupby(['Genre'])['Profit'].mean()
genre_average
sns.barplot(x=genre_average.index,y=genre_average);
- ### Subtask 3.8:. Find Books Which Earned The Most Per Year (2009-19)
data.groupby('Year')['Profit'].max()
data.groupby('Year')['Profit'].transform(max)
data[data.groupby('Year')['Profit'].transform(max) == data['Profit']]
most_earning_book_per_year=data[data.groupby('Year')['Profit'].transform(max) == data['Profit']]
most_earning_book_per_year
most_earning_book_per_year=most_earning_book_per_year.sort_values('Year').set_index('Year')
most_earning_book_per_year
- ### Subtask 3.9: What is the Average Earning of Each Genre on per Year Basis
genres_per_year_mean=data.groupby(['Year','Genre'])[['Profit']].mean().round(2)
genres_per_year_mean
- ### Subtask 3.10: Trend of books selling (Top rated) over the years
Earning_Graph=data.groupby('Year')['Profit'].sum()
Earning_Graph
sns.lineplot(data=Earning_Graph)
plt.xlabel('Year')
plt.ylabel("Earned")
plt.title("EARNING PER YEAR")
plt.figure(figsize=(12,12));
- ### Subtask 3.11: Top 10 authors which earned the most
authors=data.groupby('Author')['Profit'].sum().sort_values(ascending=False).head(10)
authors
sns.barplot(y=authors.index,x=authors)
plt.title('The Money Makers ')
