This application is built around the concept "Who has the time to read all the IMDb reviews?". This application aims to promote the wonderful films made by the Ghibli animation studio by condensing the reviews information on the IMDb website and obtaining the images and trailers about the selected movies automatically.
First the film titles and basic information are obtained using the Ghibli API. Using this information reviews were scraped from the IMDb movie reviews website. Some basic NLP techniques such as word clouds and N grams were used to summarise about 1500 user reviews. Also, the similarity between films was calculated based on the words within the reviews. Finally, the YouTube Data API was used to find the appropriate trailer for the movie automatically.
Shiny App IO: https://tigerstone.shinyapps.io/project/
https://developers.google.com/youtube/v3
Used to obtian movie trailers. Refer to YouTube.R for code.
https://ghibliapi.herokuapp.com
Use to obtain movie titles and related information such as directors and producers. Refer to ghibliPortal.R for code.
User Review data was scraped from the IMDb website and stored in the IMDB.Ghibli.Reviews table on the cloud SQL database. Refer to IMDbPortal.R for source code.
The links to the posters for the movies were obtained from the official studio ghibli website. They are stored in the table named Ghibli.Poster on the cloud SQL database. Refer to ghibliPortal.R for source code.
The data is stored on a google cloud SQL database. The database contains two basic tables as shown above. They are collected from the database and loaded in the local environment when the app starts. The database is updated automatically at 0:00 a.m. on the first day of every month via CRON job. Refer to cron.R and cron.yaml for code.
- Better UI design (for sure).
- Functions for obtaining YouTube trailer statistics and comments are implemented but not utilized.
- Obtaining more movie data and utilizing advanced NLP techniques.
