Social media platforms have long acted as sensors for monitoring the city's pulse and citizen satisfaction. Our initial studies have shown that machine learning classifiers can identify citizen issues, but turning these algorithms into a working decision-support system requires a solid framework. This work outlines the complete lifecycle of a scalable web portal designed to collect, store, and analyze Twitter data from Türkiye's Aegean Region. We explain the high-performance computing setup, the linguistic analysis of Turkish social media posts, and the launch of a public interface for real-time visualization. This paper also examines the major changes in social media research following the shift from Twitter to X. We discuss the reliability of our findings in light of the end of the free Academic Research API, highlighting how these new financial pressures create significant challenges for similar future studies. The details of the system developed in our project using open-source tools have already been documented. This work not only provides a guide to technical setup but also highlights the growing value of archived regional datasets during a period of limited data access. The results show that, despite these changing challenges, using natural language processing (NLP) and big data architecture remains essential for local administrators.
Dalkılıç et al. (Thu,) studied this question.