<h2>Description</h2> <p> This dataset is associated with the master's dissertation entitled <em>“AirCEP – An Application for Monitoring Air Quality with Complex Event Processing”</em>, developed at the Institute of Computing, University of Campinas (UNICAMP), Brazil, in 2025. </p> <h3>Study Overview</h3> <p> The study addresses the challenges of air quality monitoring in regions with limited infrastructure and high operational costs, proposing a modular and scalable architecture called <strong>AirCEP</strong>. The system integrates Edge Computing, configurable data filtering mechanisms, Complex Event Processing (CEP), and real-time visualization to reduce network traffic, computational resource consumption, and alert latency. </p> <p><strong>AirCEP was designed to:</strong></p> <ul> <li>Minimize network bandwidth usage through configurable data filters applied at the edge.</li> <li>Reduce CPU and memory consumption during real-time stream processing.</li> <li>Detect complex environmental events from continuous air quality measurements.</li> <li>Generate real-time alerts for adverse air quality conditions.</li> <li>Provide a monitoring dashboard for visualization and decision support.</li> </ul> <p><strong>The architecture is composed of three main components:</strong></p> <ol> <li> <strong>Data Router (Edge Layer):</strong> Applies configurable filters to reduce data volume before transmission. </li> <li> <strong>Stream Processing Engine:</strong> Implements Complex Event Processing using Apache Flink to analyze continuous data streams and detect patterns of interest. </li> <li> <strong>Visualization Dashboard:</strong> Built with Grafana to display real-time metrics, historical data, and alerts. </li> </ol> <h3>Dataset Description</h3> <p> The dataset includes the measurements and experimental results used to evaluate the AirCEP architecture. It contains: </p> <ul> <li><strong>Air quality measurements</strong> from monitoring stations and/or simulated sensors.</li> <li> <strong>Pollutants monitored:</strong> <ul> <li>PM2.5</li> <li>PM10</li> <li>SO₂ (Sulfur Dioxide)</li> <li>NO₂ (Nitrogen Dioxide)</li> <li>O₃ (Ozone)</li> <li>CO (Carbon Monoxide)</li> </ul> </li> <li>Timestamped readings collected as continuous data streams.</li> <li> Experimental logs from two deployment scenarios: <ul> <li><strong>Local environment:</strong> All components deployed on the same machine.</li> <li><strong>Remote environment:</strong> Sensors physically separated from the processing unit.</li> </ul> </li> <li> <strong>Performance metrics collected during experiments:</strong> <ul> <li>Network traffic (bytes transmitted)</li> <li>CPU utilization</li> <li>Memory consumption</li> <li>End-to-end latency</li> </ul> </li> <li>Filter configurations and CEP rule definitions used for event detection.</li> </ul> <h3>Data Structure</h3> <p> The dataset is structured to support reproducibility and reuse: </p> <ul> <li>Time-series formatted records (timestamp + pollutant measurements).</li> <li>System performance logs aligned with experimental runs.</li> <li>Configuration files defining filtering thresholds and CEP rules.</li> <li>Scenario identification metadata (local vs. remote deployment).</li> </ul> <h3>Methodology Context</h3> <p>The experiments compare:</p> <ul> <li>Baseline transmission (without filtering) versus filtered transmission.</li> <li>Resource consumption in different deployment architectures.</li> <li>Latency impact of physical distance between sensors and processing nodes.</li> </ul> <p> Results demonstrate: </p> <ul> <li>Up to ~30% reduction in network traffic.</li> <li>Up to ~19% reduction in CPU consumption.</li> <li>Latency more strongly influenced by physical distance than filtering itself.</li> </ul> <h3>Reuse Potential</h3> <p> This dataset can be reused for: </p> <ul> <li>Research in real-time stream processing.</li> <li>Complex Event Processing evaluation.</li> <li>Edge Computing performance studies.</li> <li>Network optimization experiments.</li> <li>Air quality monitoring systems benchmarking.</li> <li>Smart city and IoT research.</li> <li>Comparative analysis between cloud-centric and edge-based architectures.</li> </ul> <p> Researchers may replicate the experimental setup, validate performance trade-offs, benchmark alternative stream processing engines, or test new event detection rules. </p>