simeongerginov/github-recommender

GitHub-Recommender

github-recommender's Issues

[Task] Define data structure for the Recommender System

The structure of the data needs to be defined, before the GitHub crawler is implemented. My proposition is to have the following datasets:

1. users.csv containing the following columns: Id, Name and Type. The Type column for the user specifies whether the user is a User or an Organization.
2. repositories.csv containing the following columns: Id, Name, Description, Languages, OwnerId, CreatedAt and UpdatedAt.
3. activities.csv containing the following columns: RepositoryId, UserId and ActivityType. ActivityType can be one of the following: Watch, Star and Fork.

@PacoPacov and @Jor0Stam, do you think any additional columns need to be added ?

[Task] Research Deep Learning for Recommender Systems

Research the Deep Learning for Recommender Systems approach and if applicable, use it for recommending GitHub repositories to a user.

[Task] Use Non Personalised and Stereotyped Recommendation approach on the GitHub dataset

On the created GitHub dataset, apply the Non Personalised and Stereotyped Recommendation approach for recommending GitHub repositories to a user.

[Task] Use Content Based Recommendation (CBR) approach on the GitHub dataset

On the created GitHub dataset, apply the Content Based Recommendation (CBR) approach for recommending GitHub repositories to a user. Try out the different types of content based filtering and compare the results.

[Task] Use Item Item Collaborative Filtering approach on the GitHub dataset

On the created GitHub dataset, apply the Item Item Collaborative Filtering approach for recommending GitHub repositories to a user. Try out the different approaches and compare the results.

[Task] Research Hybrid Recommenders

Research the Hybrid Recommenders approach and if applicable, use it for recommending GitHub repositories to a user.

[Task] Implement GitHub data crawler using the GitHub API

The datasets that are needed for the GitHub Recommender need to be created from the data provided via the GitHub API.

[Task] Use User User Collaborative Filtering approach on the GitHub dataset

On the created GitHub dataset, apply the User User Collaborative Filtering approach for recommending GitHub repositories to a user. Try out the different approaches and compare the results.