Predicting Movie Rental Durations

A DVD rental company needs your help! They want to figure out how many days a customer will rent a DVD for based on some features and has approached you for help. They want you to try out some regression models which will help predict the number of days a customer will rent a DVD for. The company wants a model which yeilds a MSE of 3 or less on a test set. The model you make will help the company become more efficient inventory planning.

The data they provided is in the csv file rental_info.csv. It has the following features:

"rental_date": The date (and time) the customer rents the DVD.
"return_date": The date (and time) the customer returns the DVD.
"amount": The amount paid by the customer for renting the DVD.
"amount_2": The square of "amount".
"rental_rate": The rate at which the DVD is rented for.
"rental_rate_2": The square of "rental_rate".
"release_year": The year the movie being rented was released.
"length": Lenght of the movie being rented, in minuites.
"length_2": The square of "length".
"replacement_cost": The amount it will cost the company to replace the DVD.
"special_features": Any special features, for example trailers/deleted scenes that the DVD also has.
"NC-17", "PG", "PG-13", "R": These columns are dummy variables of the rating of the movie. It takes the value 1 if the move is rated as the column name and 0 otherwise. For your convinience, the reference dummy has already been dropped.

1 hidden cell

How to approach this project

Getting the number of rental days.
Adding dummy variables using the special features column.
Executing a train-test split
Performing feature selection
Choosing models and performing hyperparameter tuning
Predicting values on test set
Computing mean squared error

import pandas as pd
import numpy as np

from sklearn.model_selection import train_test_split
from sklearn.metrics import mean_squared_error

# Import any additional modules and start coding below

# Investigate the dataset
df = pd.read_csv('rental_info.csv')

display(df.head())
display(df.info())
display(df.describe())
display(df.isna().sum())

# 1. Getting the number of rental days.
# Convert the columns to datetime
df['rental_date'] = pd.to_datetime(df['rental_date'], format='%Y-%m-%d %H:%M:%S%z', utc=True)
df['return_date'] = pd.to_datetime(df['return_date'], format='%Y-%m-%d %H:%M:%S%z', utc=True)

# Create a new column 'rental_length_days'
df['rental_length_days'] = df['return_date'] - df['rental_date']

# Select only days, excluding other components.
df['rental_length_days'] = df['rental_length_days'].dt.days

# Create two columns of dummy variables from "special_features", which takes the value of 1 when:
# The value is "Deleted Scenes", storing as a column called "deleted_scenes".
# The value is "Behind the Scenes", storing as a column called "behind_the_scenes".

df['deleted_scenes'] = np.where(df['special_features'].str.contains('Deleted Scenes'), 1,0)
df['behind_the_scenes'] = np.where(df['special_features'].str.contains('Behind the Scenes'), 1,0)

# Executing a train-test split
#Split the data into train and test sets, avoiding any features that leak data about the target variable, and include 20% of the total data in the test set.

# Decide which columns to use
cols_to_drop =["special_features", "rental_length_days", "rental_date", "return_date"]
X = df.drop(cols_to_drop, axis=1)
y = df['rental_length_days'] 

# Split data into train & test
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=9)

# Performing feature selection
# Use Lasso model - why?
# For lasso
from sklearn.linear_model import Lasso
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler

# Create the Lasso model
lasso = Lasso(alpha=0.3, random_state=9) 

# Train the model and access the coefficients
lasso.fit(X_train, y_train)
lasso_coef = lasso.coef_

# Perform feature selectino by choosing columns with positive coefficients
X_lasso_train, X_lasso_test = X_train.iloc[:, lasso_coef > 0], X_test.iloc[:, lasso_coef > 0]

# Choosing models and performing hyperparameter tuning - Try a variety of regression models.

# Run OLS ("Ordinary Least Squares")
from sklearn.linear_model import LinearRegression
from sklearn.metrics import mean_squared_error

ols = LinearRegression()
ols = ols.fit(X_lasso_train, y_train)
y_test_pred = ols.predict(X_lasso_test)
mse_lin_reg_lasso = mean_squared_error(y_test, y_test_pred)


# Random forest
from sklearn.ensemble import RandomForestRegressor
from sklearn.model_selection import RandomizedSearchCV

# Random forest hyperparameter space
param_dist = {'n_estimators': np.arange(1,101,1),
          'max_depth':np.arange(1,11,1)}

# Create a random forest regressor
rf = RandomForestRegressor()

# Use random search to find the best hyperparameters
rand_search = RandomizedSearchCV(rf, 
                                 param_distributions=param_dist, 
                                 cv=5, 
                                 random_state=9)

# Fit the random search object to the data
rand_search.fit(X_train, y_train)

# Create a variable for the best hyper param
hyper_params = rand_search.best_params_

# {'n_estimators': 51, 'max_depth': 10}

# Run the random forest on the chosen hyper parameters
rf = RandomForestRegressor(n_estimators=hyper_params["n_estimators"], 
                           max_depth=hyper_params["max_depth"], 
                           random_state=9)

rf.fit(X_train,y_train)
rf_pred = rf.predict(X_test)
mse_random_forest= mean_squared_error(y_test, rf_pred)

# Random forest gives lowest MSE so:
best_model = rf
best_mse = mse_random_forest

# Print and check the dataset
display(df['rental_length_days'])
display(df)
mse_lin_reg_lasso
hyper_params
display(best_model)
display(best_mse)
display(mse_lin_reg_lasso)
display(mse_random_forest)