Machine Learning Project
At first we had 3 different datasets (csv files); movie-director.csv, movies-revenue.csv and movie-voice-actors.csv which we had to merge in a single dataset.
Contains about 45 movies where each movie has its title and director.
Contains about 463 movies where each movie has its title, release date, genre, MPAA rating and revenue.
Contains about 811 records of characters with their voice actors and featured movies.
First, we have merged the three dataframes that read the three csv files into a new dataframe "movies_df". Then, we have applied changes on the new dataframe by calling data_formatting function whose functionality will be discussed later. Then, the following preprocessing techniques; feature scaling, ordinal encoding & one hot encoding, have been applied on the dataframe using functions we have defined.
We pass "movie_df" dataframe to it as a parameter to apply some changes on its columns.
- "revenue" column:
-"$" and "," symbols are removed from all revenues
-the records are numerized (where it is possible) - "release_date" column:
-all null values are replaced by zeros
-its format is changed from the following for example:12-JAN-88 to the year only 1988
-its type is converted to integer (int32)
At the end after applying the changes the data frame is returned
It's called by passing "movie_df" data frame to it as a paramter to change the scale of numbers for each column to be between 0 and 1 all inclusive.