Naagarik: A Machine Learning Framework for Intelligent Analysis of Civic Issues
Adithi Satish, Shriya B Shankar, K N Kavitha · 2021 Asian Conference on Innovation in Technology (ASIANCON) · 2021
This paper introduces Naagarik, a framework which analyses hyperlocal, real-time citizen-generated tweets filtered by location, and determines whether they are civic issues or not. It further classifies the civic issues into different categories like Waste/Garbage, Water, Potholes and more. In addition to this, we compute the sentiment of the reports to capture their severity. We use Machine Learning and Natural Language Processing techniques like Logistic Regression, Support Vector Machines and the VADER sentiment analyzer to perform the three-step process of identification, classification and sentiment analysis respectively. These models are chosen on the basis of better performance metrics exhibited, when compared to other text classification algorithms. Thus, our framework bridges the gap between a city's local government and its citizens by highlighting their grievances based on relevance, category and severity.