thejas-bhat/biasdetection

Bias Detection is a project that analyzes political tweets to find which users hold the most influence in a social network, and what topics people are talking about. Think of it as mapping out who drives conversations in political discussions on Twitter. The project works in two parts. First, it builds a network graph of Twitter users and runs something called "Critical Node Detection" to pinpoint the most influential people in that network. It offers several versions of this analysis, including ones that use parallel computing through OpenCL to speed things up. Second, it performs topic modeling on the tweets themselves, which means grouping tweets by subject matter to understand what themes dominate the conversation. This would be useful for researchers, journalists, or campaign strategists who want to understand how political information spreads on social media. For example, a news organization could use it to identify which accounts shape public opinion during an election, or a research team could study how biased narratives form around specific topics. The project is written primarily in Python but requires a C++17 compiler and CMake to build, since the performance-critical network analysis runs in C++. The README doesn't go into detail about how the tweet dataset is structured or what the topic modeling output looks like, so you would need to dig into the code itself to understand those specifics. It's worth noting that this appears to be an academic or research-oriented project rather than a polished product. The README focuses on build instructions rather than explaining how to feed in your own data, so using it would likely require some comfort with development tools.