I first got involved with research after my freshman year, when I applied and was accepted into the Kansas Louis Stokes Alliance for Minority Participation (KS-LSAMP) Research Immersion: Pathways to STEM (RiPS) program. After that, I officially joined K-State's Laboratory for Knowledge Discovery in Databases (KDD) as an undergraduate research programmer.
Summer 2019: KS-LSAMP RiPS
During this summer, I became fully immersed into the world of research. I began working on a project named Threat Intelligence, which is focused on automatically identifying cyber threats using machine learning techniques on datasets gathered from social media platforms. After reviewing literature and publications for the project, I spent a lot of time learning the basics of machine learning, cyber security, and social network analysis. This included learning how to train convolutional neural network models for binary and multi-class classification of Twitter data that was previously annotated for the project.
After some initial learning and getting caught up to speed with the project, I began working to scrape new data from Twitter using Python, and then applied cloud-based natural language processing APIs to extract more information. This included analyzing the sentiment of each tweet and applying topic modeling to extract keywords and entities from the tweets. I was then able to take that information and visualize it using R.
To showcase all of my work at the end of the summer, I wrote an abstract for the research I performed and presented it at a final poster session for the research program. It was so great to be able to share my work with those who attended the poster session, but was even more interesting to see what other students in the program worked on as well.
Fall 2019: Joining K-State KDD
After the summer, I received an offer to officially join KDD as an undergraduate research programmer. I continued to work in collaboration with the University of New Haven on the Threat Intelligence team, with a focus on information extraction and information visualization.
Currently, I am working with Elasticsearch and Kibana to visualize our currently annotated data. Elasticsearch provides various searching and data aggregation capabilities, while Kibana gives us a way to visualize all of our data in various formats, as well as drill-down to fine details.