Data and source code for a badge recommendation experiment
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.

README.md

Data and source code for the paper "Gamified Incentives: A Badge Recommendation Model to Improve User Engagement in Social Networking Websites". ( pdf )

Repository Description:

data:

Contains badges dataset for different time periods which are extracted from Badges.xml data of Stack Overflow using dataset_generation/extract_badges.py.

  • dataset2008: Contains randomly generated train and test dataset from badges dataset using dataset_generation/train_test_generation.py for badges that are awarded in the year 2008. The complete dataset for years 2008 to 2010 are compressed in the datasets.zip file and should be uncompressed like dataset2008 to be used.

dataset_generation:

  • extract_badges.py: Extracts badges for users which are awarded within the begin_year and end_year time period from the Badges.xml file. It writes the output as a csv file in the format of UserId,badge1,badg2,...
  • train_test_generation.py: Generates train and test set from badges.csv file for recommendation/ algorithms.

recommendation:

  • collaborative_filtering.py: Implementation of the item-based collaborative filtering method to recommend badges and evaluate the results.
  • popular_badge_baseline.py: Implementation of the baseline algorithm which recommends popular badges to each user.
  • datasets.py: This module contains the file paths of datasets for easier access in other modules. If you change the repository structure and the place of the data directory, you should modify _DATASET_ROOT in this module accordingly so that it points to the root directory where datasets reside in.