055

Title:

Building a Framework of Software and Hardware for Real-time Analytics on Social Networks

Discipline: Computer Science

Presenter:

Mehtab Sidhu

Abstract:

Facebook reported that in the second quarter of 2019 they had over 2.4 billion monthly active users, an 8% increase year-over-year. These large data-sets are constantly changing as new users join or as users make new posts or even interact with existing posts. These changes reflect the current state of the mindsets and behaviors of users, organizations, institutions, and even countries. Capturing major changes is critical to the success of social networks. These data-sets hold a goldmine of information that can be used for generating trends, networks,relationships, forecasting, preferences, etc. These huge data-sets can be modeled for analysis using a Graph - a collection of entities called nodes or vertices connected to each other by links called edges. In a Facebook graph, nodes represent individual objects such as a user, a photo, a page, or a comment, and edges represent connections between a collection of objects and a single object, such as photos on a page, or comments on a photo.

Capturing major changes to large data-sets in real-time presents a challenge. We are building a framework to demonstrate that dynamic spectral graph partitioning can enable real-time analytics on large-scale social networks. The long term objective of this research is to re-partition dynamically changing graphs to capture current changes in real-time and build a framework of hardware and software that can capture changes in social network graphs in real-time.

Author(s):

Mehtab Sidhu, Andrew Sohn

Funding Acknowledgements:

nan