Project 2: Regression - Adapting to Succeed?

Weeks 2 and 3

Backstory:

Using information we scrape from the web, build linear regression models from which we can learn about movies adapted from novels and predict whether an adaptaion will success.

Data:

• acquisition: web scraping
• storage: flat files
• sources:

• movie: IMDb
• book: goodreads, fictionDB, Google

Target variable:

Opening weekend box office (U.S.)

Features:

Movie

• MPAA rating
• release year
• release seaon
• director
  • number of directed films then
  • average rating of directed films then
  • average gross then
• cast
  • number of participated films then
  • average rating of participated films then
  • average gross then
• writer: with original author or not
• genre
• distributor
• language
• country

Book

• first published year
• years from movie release
• page
• author
  • number of publications then
  • visibility
• book visibility

Skills:

• basics of the web (requests, HTML, CSS, JavaScript)
• web scraping (BeautifulSouop, selenium)
• numpy and pandas
• statsmodels, scikit-learn
• matplotlib, seaborn

Analysis:

• Model: Linear Regression
• Featrue engineering: StandardScaler,PolynomialFeatures,OneHotEncoder
• Cross-Validation
• Regularization: Ridge,Lasso,RidgeCV,LassoCV