Can Search Engine Optimization (SEO) signals be used to detect unique structural aspects of disinformation sites?

What is this project?

Barn Owl investigates if disinformation sites can be correctly identified from a group of misinformation and neutral sites through a combination of SEO signals. We investigate the following by combining SEO marketing tools and Python:

A.) Useful disinformation data can be pulled independently with marketing techniques - no need for potential filtering by third parties

B.) Small and Mid-sized disinformation sites can structurally be detected separately from misinformation sites.

• If so, what combination of structures work?
• What is their accuracy?

C.) Small and Mid-sized disinformation sites can be detected on the backside without need of content analysis

D.) Small and Mid-sized disinformation sites have a unique set of SEO signal identifiers and tendencies

E.) If there is a unique structure, is it applicable across multiple languages?

History of this project

Goals for the semester

• Grow existing data’s neutral, misinformation, and disinformation list
• Use Pandas to explore small sample set
• Write findings
• Apply research to larger dataset —> Possibly partner with ARL:UT for Machine Learning (ML)

What students will learn?

• Search Engine Optimization (SEO)
• Network Analysis
• Giphy
• Python (Pandas)
• Freedom of Speech Policy
• Data Visualization
• Research development and process
• Scientific Writing

Deliverables and Goals

• Write up of small exploratory findings
• Pandas code available to GitHub for exploratory analysis
• Present to Sandia for renewal
• Completed dataset with seed sites and backlinks data
• Expand larger dataset to include foreign language sample and compare findings