Capstone_Project

Capstone Project IBDM Dataset

This capstone is divided into a 4-phase duration. In this document, the entire details of the 4-phase(weekly) duration are given.

Phase 1

1. Webscrape the provided URL - IMDB dataset:

https://www.imdb.com/search/title/?genres=action&sort=user_rating,desc&title_type=feature&num_votes=25000,&pf_rd_m=A2FGELUUNOQJNL&pf_rd_p=f11158cc-b50b-4c4d-b0a2-40b32863395b&pf_rd_r=XZ8X52H1R40B7KG5SNZ9&pf_rd_s=right-6&pf_rd_t=15506&pf_rd_i=top&ref_=chttp_gnr_1

2. Store the entire data in two different CSV files as per the given fields:

The first CSV file data contains :

Sno, Movie Name, Director Name, Duration, year, ratings, Metascore

Bifurcate the Director field into subfields as per the number of directors of the movie belongs to such as Director1, director2

The second CSV file contains the following:

Movie Name, stars, votes, Genre, Gross collection, popularity, Certification

Bifurcate the stars field into 4 subfields as per the number of stars worked in the movie such as star1, star2, star3, star4

Bifurcate the genre into 3 subfields as per the number of genres the movie belongs to such as :

Genre1, genre2, genre3

Phase 2

Make two tables and corresponding columns provided in the above CSV files in SQLite DBMS. Insert all data of each CSV file in each of the created tables. Now start querying the table(s) in the SQL workbench / SQLite database :

Table 1: Sno, MovieName, Director Name, Duration, genre, ratings

1. Display all the details of movies created by directors Christopher and Matt Reeves.

2. Display all the details of movies with a duration of 140 minutes to 190 minutes.

3. Display all details of movies with ratings above 7 in ascending order.

4. Display all movie names in descending order.

5. Display movie name starts with ‘P’ and their rating is greater than 7.

Table 2: Movie Name, stars, votes, Genre, Gross collection, popularity, Certification

1. Display all movie names with star Arnold Schwarzenegger in ascending order.

2 2) Display all details of the movie with the highest number of votes.

3. Display movie names with gross collections in descending order.

4. Display the gross collection of movies with the star Arnold.

5. Display all details of movies with comedy and action genres.

Make subquery :

1. Display all details from both tables where movie names are the same.

2. Display all movie names, Director, ratings, and gross collection where the genre is action.

3.  Display all details from both tables with the highest gross collection.
   

4.  Display all details from both tables with the highest ratings
   

5.  Display all details from both tables with the lowest gross collection and lowest ratings
   

Now once completed with queries in the SQLite database, then make the exact query solutions by using PANDAS SQL in the data frame. Load the CSV data in a data frame and start making solutions for all the above 15 queries using PANDAS SQL. You may use concat or merge joins per the requirements basis to make 5 join queries.

Phase 3

1. Now make only 1 data frame of two CSV files using the join operation of pandas and start doing EDA.

2. Do the complete EDA in detail to explore the insights of data and write detailed observations of each analysis.

Phase 4

1. Write the complete Machine learning code to make predictions of votes and gross collection. Use appropriate models on their label basis. Remember you need to make 2 different predictions: vote and gross collection.

2. Apply all the best techniques of scaling, and hyperparameter tuning, and avoid underfitting or overfitting (bias/variance)

3. At the end save the best model and convey on which basis you have chosen that model.