Mapping Makassar’s Danger Zones: How Data Science is Making Our Roads Safer

GOWA, Indonesia June 19 2026/CDM/ – Every single day, millions of people navigate city streets, completely unaware of the hidden hazards around the next corner. According to the World Health Organization, traffic accidents remain the leading cause of death globally for children and young adults. This pressing crisis directly intersects with the United Nations’ Sustainable Development Goals, specifically SDGs 3 (Good Health and Well-Being), which sets an urgent milestone to drastically minimize global traffic mortalities. To combat this, a team of forward-thinking researchers from Universitas Hasanuddin published a breakthrough study in April 2025 titled “IDENTIFICATION OF TRAFFIC ACCIDENT PRONE AREAS USING DATA MINING” within the journal ICIC Express Letters, published by ICIC International. Their mission? Using advanced data analytics to accurately pinpoint accident hotspots and ultimately save human lives.

The researchers turned their attention to the bustling streets of Makassar City, Indonesia, analyzing a massive dataset of 3,558 historic daily traffic accident records. Instead of manually guessing where accidents occur most frequently, they deployed an intelligent machine learning system powered by the DBSCAN (Density-Based Spatial Clustering of Application with Noise) algorithm. This smart algorithm acts like a digital detective, automatically grouping massive amounts of scattershot location data into highly dense, tight-knit geographic clusters based on real-world similarities.

To make the algorithm incredibly precise, the team mathematically streamlined their data using Principal Component Analysis (PCA) to remove irrelevant noise and clutter. Through careful hyperparameter tuning, their system successfully mapped out 16 distinct, high-risk traffic accident clusters across Makassar. When compared against traditional, less flexible grouping methods like the K-Means algorithm, DBSCAN proved far superior. It achieved a remarkably high Silhouette Score of 0.805 and a low Sum Square Error of 71.22, meaning the identified danger zones were incredibly accurate and sharply defined.

The magic happens when this abstract math is transformed into something anyone can use. The study seamlessly integrated these 16 data clusters into an interactive WebGIS dashboard—a dynamic digital map equipped with clever tooltips. By clicking on a marked hotspot, city officials, traffic police, and everyday commuters can immediately view detailed breakdowns of past incidents. This includes typical accident types—ranging from head-on collisions to hit-and-runs—and the specific vehicles involved, such as motorcycles or passenger cars.

By turning thousands of chaotic paper records into a living visual map, data science provides an actionable blueprint for urban planning. Local authorities can now deploy precise preventative measures, fix poor road infrastructure, and strategically allocate emergency medical services right where they are needed most. Ultimately, this innovative fusion of data mining and digital mapping proves that tech-driven solutions are vital to fulfilling SDGs 3, ensuring our shared public roads transition from dangerous hazards into secure, sustainable pathways for everyone.