Movies Recommendation Engine

Built a Movie Recommendation Engine in the form of a web-app which recommends movies using user-user collaborative filtering.

Tech Stack

Built API using web2py (Python based framework) and Front-end using JavaScript with VueJS

Installation Instructions

Clone the repository and run:

python web2py.py

That's it!!!

Directory structure

project/
    README
    LICENSE
    VERSION                    > this web2py version
    web2py.py                  > the startup script
    anyserver.py               > to run with third party servers
    ...                        > other handlers and example files
    gluon/                     > the core libraries
        packages/              > web2py submodules
          dal/
        contrib/               > third party libraries
        tests/                 > unittests
    applications/              > are the apps
        admin/                 > web based IDE
            ...
        examples/              > examples, docs, links
            ...
        welcome/               > the scaffolding app (they all copy it)
            ABOUT
            LICENSE
            models/
            views/
            controllers/
            sessions/
            errors/
            cache/
            static/
            uploads/
            modules/
            cron/
            tests/
        ...                    > your own apps
    examples/                  > example config files, mv .. and customize
    extras/                    > other files which are required for building web2py
    scripts/                   > utility and installation scripts
    handlers/
        wsgihandler.py         > handler to connect to WSGI
        ...                    > handlers for Fast-CGI, SCGI, Gevent, etc
    site-packages/             > additional optional modules
    logs/                      > log files will go in there
    deposit/                   > a place where web2py stores apps temporarily

Recommendations Model File:

item_similarity_df.csv

Posters for each movie is from MovieLens dataset. Filename corresponds to the ‘movieId’. It is open-sourced at: https://github.com/babu-thomas/movielens-poster