Gaining insights from Question/Answering systems using Graph Algorithms

[d_boukraa]

Hi everybody
this project is about applying Neo4j graph algorithms (or the GDS library) to gain insights from question / answering systems such as StackExchange (SE). Examples of insights include (but are not limited to ) ranking users (using PageRank algorithm), personalised ranking of users, (using PPR) and community detection.
The project is at its beginning: we have already imported data from SE Network about users and their interactions as CSV files. We have already modelled a first type of user interaction as (user) -[:comments] -> (user) in Neo4j graph and we are currently working on modelling other user interactions (votes, answers)...

Our project can be applied to any SE sub-network such as StackOverflow, AskUbuntu, etc. since all the sub-networks share the same database schema.

We hope the discovered knowledge will help enhance user interactions in Q/A systems, for instance by recommending posts according to user profiles, unveiling existing communities, etc.

We will create a related GitHub project soon.

[sameerG]

Hi Douikifli
Is above mentioned project ready.Can I install it for preview?

[d_boukraa]

Hi Sameer,

sorry for my late reply. We have just completed the project. I'll post details about the application we developed and the way to use it.

Regards,

Doulkifli

[sameerG]

Hi Doulkifli

Do you have the GitHub url so that I can review it and see if it generates any useful insights
Regards
Sameer

[doulkifli]

Hi Sameer,

Sorry for my late reply. We completed the project a couple of months ago. Indeed, I have a GitHub url but I have to add the projects files. Once added, I'll inform you.

Regards,

Doulkifli

[d_boukraa]

Hi Sameer, I've just added the project files in my GitHub Space. You can access it here

github.com

Doulkifli/Question-AnsweringGraphs

This repository contains the CSV files and the Cypher queries of the project