Social Analytics Project

• An academic project for social analytics.

Directories Summary

Lets rock

• code - this is the directory for all codes.
• docs - used for .github.io/saproject presentation purposes
• references - cheatsheets and references
• report - proposal and report in md format

Introduction

This project is for our social analytics course. In this project, we built a telegram chatbot to rocommend restaurants and cuisines in Sngapore using social graphs. We got the data we use from FourSquare API.

Methodology

1. Prepare dataset
2. Build kmeans clustering model to classify geographic locations of users.
3. Build GMM (Kmeans) clustering model to classify food preference of users.
4. Build Spectral clusterings to model user-restaurant, user-user relationship.
5. Recommend based on user profiles and (or, for new user) user relationships.
6. Interact with user through telegram jiak_bot.

Result

See this report for more information about our project.
https://github.com/yc930401/recommendation-engine-using-social-graph/blob/master/report/Grp5_ProjectReport.pdf