Film Findr – Eric Marshall

Background

There are dozens of movies released every year – far more than any single person (or even a group of people) can expect to keep up with. While there are websites such as IMDB that allow users to search for facts and information about movies, these sites do not also look for relationships between movies. Having access to such a database would provide both academics and film enthusiasts with a new resource for examining connections between movies – from the words in similar scenes to the colors that were used across a movie.

In such a system, a key feature is the ability to search and navigate easily. By including many possible relations for users to search on, we can guarantee that there will be a wealth of information that becomes available to them. Similarly, making it easy for users to navigate through the database (and even upload films that they wish to be able to analyze) will make the database more accessible and applicable to everyone. Both of these features will make the online database invaluable to anyone who has the desire to examine connections and parallels between different films.

Executive Summary

This project has already been started as a web application and that is certainly the best medium for it, due to the fact that it will provide the most people access on the most devices. At the moment, the data and search functionality are displayed through simple web pages, which could be brought more to life with a reactive interface that would better present the information pertinent to a user’s search, while still allowing them to view more information if desired.

There should also be a feature for uploading new movies, which would allow for the server to automatically pull data from movie files (screenshots, subtitles, etc) and could even have the images taken analyzed by a computer-vision program to retrieve information such as objects in the image from them, allowing for more meaningful searches and relational matches.

Viability Analysis

As with any cloud-hosted database, the cost of hosting will increase with the number of movies that are held in it. Thus, there will be cost constraints associated with the number of movies that can be held in the database while operating within a reasonable budget. Another constraint will come from any sort of remote library for image analysis. The majority of libraries that will be able to successfully analyze the images for metadata in the database will have some cost associated with them (see the Google Vision API as an example). Finally, for uploading we may also wish to consider creating a local application for parsing movies so the size of files that have to be uploaded are drastically reduced in size.

Risks and Rewards

This project has one primary legal exposure with it, involving the use of movies. The movie industry is known for aggressively defending its copyrights and going after people for violating them. Because the database uses only screenshots, however, this should not be an issue with fair use laws. The benefits, on the other hand are significant. Having access to a massive relational database of movies with extra metadata associated with them will allow film buffs and scholars to make connections where they might have not before.

Closing

The film database that is being created will certainly be a revolutionary tool for scholars and film buffs everywhere. To make it truly useful, however, it needs to have comprehensive search and uploading features that are accessible through a friendly and reactive interface, all of which this project will add to the database.

Print PDF