Review of books from the dataset Goodreads-books

Reading improves memory and cognitive ability — which means its a healthy habit that can help kids in and outside the classroom. Along with that improved brain power comes an improved vocabulary.

“Come to the book as you would come to an unexplored land.
Come without a map. Explore it and draw your own map.” 
                                                                                          – Stephen King 

“The only thing that you absolutely have to know, is the location of the library.” 
                                                                                          – Albert Einstein

So, I don't feel I need to give a long talk to convince you the relevance of the project or problem. Reading books are compulsory. But now a days soft copy of many books are freely available. So, publishers have to have the updated information about the market, about the demand and choice. Here by this dataset one can draw some meaningful inferences from it which might progress the publishers economic growth.

About the dataset:

We have the following features or culumns with us after modification: bookID: 11127 distinct values title: 10352 distinct values authors: 6643 distinct values average rating: 209 distinct values isbn: 11127 distinct values isbn13: 23 distinct values language code: 27 distinct values no of pages: 997 distinct values ratings count: 5294 distinct values text reviews count: 1822 distinct values publication date: 87 distinct values publisher: 2292 distinct values rating: 4 distinct values Here is the description,

Detailed explaination and some inferences from the analysis:

• (1) Distribution of average rating: From the distribution of the average rating it is clear that almost all the books have been rated in the range 3.5- 4.5 . Out of almost 11000 books 22 books is rated as ‘’5’’.

• (2)Dependencies between different review parameters and average ratings: We can see that maximum books got reviews less than 20000

• (3) . Number of books in different languages: Here maximum books are written in “eng” language . 80% book is written in “eng” 13% in “en-US” and the remaining 7% is in other language.

• (4) Number of books in different page ranges: Pages of maximum books(around 60%) are in the range of 200-400
• (5) Number of authors having exact number of books: There are more than 6000 authors who wrote only one book . And 1000 authors wrote two books . The number of authors who wrote more than 5 books is very low.
• (6) Books published in a different years: It is seen that in every year no of published books is increased largely . But after 2016 it suddenly goes down.

• (7)Dependency between Average rating and different languages: Though the number of books written in “eng” is much greater than other languages , average rating of books written in 'eng' is ~3.9 . Whereas very few books are written in “wel” but the average rating is 5 .

• (8) Relation between Average Rating and Number of pages: We can find here that rating is almost varies as the num of pages of the book . Book with no of pages 2600-2700 is of high rated . Though there are few books in this range , It actually ensures that if a book is interesting to the readers, total number pages of book does not matter.

• (9)Dependencies of the various review parameters with average ratings: As we earlier saw that there are very few books with very high number of reviews . The books with very high reviews also fall under the rating range 3.5-4.5 . So, we can say maximum people, who gives a review , responds to a book with an average rating of 3.5-4.5 ,unless they get a good satisfaction with the book or become totally hopeless .Which is intuitively acceptable.

• (10) How average rating is differed for different publishers: There are 2290 publishers published the books . Out of 2290 publishers 47 publishers got average rating greater than 4.5 , 2190 got in the range [3.5,4.5] , and the remaining 134 publishers got rating less than 3.5. the approximate view is clear from the corresponding plot. See ipynb files for the plot.

• (11) Average number of pages for different languages: Among the languages, the most used language is ‘eng ’ . And we can see that books with “eng” language have 300 pages in average . Whereas books written in “enm” has maximum no of average pages (around 1100 ) . But it is also a point to note that books of most average- rated language “wel” are of total average pages 150 ( which is minimum among the used languages )

• (12) Number of reviews for different languages: As the most used language to write a book is “eng”, it has highest total reviews . But also it has a highest number of reviews which is far more from others . So we can conclude that most of the people who read books of ‘eng’ language used to give reviews after reading it .

• (13) Dependency between Number of pages and Review parameters: Books are reviewed most for the books having pages less than 1000. Number of reviews for books having more number of pages are very low.

• (14) Comparison between top three authors having most number of books: • Initially averae rating decreases fore more number of pages ,but later on it increases for all 3 author. • For low number of reviews all the authors has a mixed average rating. But when total number of reviews increases, average rating decreased for J.R.R. Tolkein and William Shakespeare but increases for Stephan King. • All three author has more reviews in pages between 250 to 750.

• (15) Comparison between top three publishers: • For higher number of reviews, average ratings are below 4 for these publishers • Average rating has not such pattern for different number of pages • For hiher number of paes , total reviews decreases for all publishers

About this repo: (How to run the files):

Here is the short instruction to understand the flow of this reprosetory

There in the repo you will find PLOTS_R which means all plottings in Rstudio using R code and lybraries. plots_python is similarly all plootings of data in Jupytor Notebook using Python lybraries Mahendra_nandi_P and Mahendra_nandi_R here P means pyhton and R means Rstudio so .ipynb files are created Jupytor Ntebook and .R files are created in Rstudio You will find the books.csv file which is the required file for data and other 3 .csv files are of no use here though . The final reports are the .pdf and .pptx files . So you will eventually get every thing inside it.

• NOTE: I have perhaps changed the books.csv file mannually a little bit so the modified file naming booksP.csv will be uploaded here too. And I have collected the data from kuggle. HERE YOU can find IT

Some sample Visuals:

Visualization using R:

• .1
• .2
• .3
• .4
• .5
• .5

Visualization using PYTHON:

• .77

• .55

• .6

• .7

• .1

• .2

• .3

• .4

• .8

• .33

• .11

• .22

• .66

• .9

• .0

• .44

• .5

The further task is to build a prediction model which will make the work colplete. Thank You . If you have any querry about anything please feel free to contact and also for any suggestion..

here is my email ID : [email protected] My LimkedIn profile : Mahendra Nandi

Asif Ikbal was there with me in The first part of the project in R. Thank you Asif.