Emergency Demand Forecasting

[Audio] Emergency Demand Forecasting and Forecast-Informed Ambulance Positioning A Case Study of the e-Dispatch System of Silang MDRRMO Preliminary Research Concept Evidence Brief NEO · GROUP D · OCTOBER 30, 2026.

[Audio] Research problem & significance Demand moves. Ambulances cannot be everywhere. UNDERSTAND ANTICIPATE hotspots POSITION resources demand patterns EMS demand varies across time and place, complicating deployment. Historical e-Dispatch records may reveal patterns—but their practical value for Silang MDRRMO remains uncertain. Potential contribution Predictive insight → prescriptive action.

[Audio] What current research tells us Forecasting is established. Fine-grained reliability is not. MODELS GIS EVIDENCE OPEN CHALLENGE 23.01% Sparse + zero Gradient Boosting + AdaBoost Portuguese study highlighted strong machine-learning approaches. Best daily MAPE reported by Chen et al. (2016). Low counts and zero-demand periods weaken fine-grained forecasts..

[Audio] The gap is not another algorithm. It is local proof that forecasts improve operational decisions under Silang's real constraints. VALIDATE LOCALLY / ASSESS OPERATIONALLY.

[Audio] Five questions shape the study 01 02 03 What patterns emerge across time and geography? How accurately can history predict future demand? Which model performs best with local data? 04 05 How should forecasts guide positioning within fleet and geography? Does forecast-informed positioning outperform current deployment?.

[Audio] Expected contribution & NIBRA alignment Disaster risk reduction Community safety & management & preparedness A locally validated forecasting + ambulance positioning framework Data science, AI Resilient, efficient & digital innovation public service.

[Audio] From prediction to positioning A sequential predictive–prescriptive framework Integrate Explore incidents Translate forecasts decision support Compare models Evaluate strategies Time · location · category.

[Audio] Feasibility depends on three linked realities INPUTS CONSTRAINTS EVALUATION e-Dispatch history Fleet + deployment data GIS + road network Operational rules Fleet availability MDRRMO coordination Demand coverage Travel time Resource utilization.

[Audio] Responsible AI means useful, not novel EMS forecasting algorithms are established. The research value lies in evidence-based, context-sensitive application—and transparent limits. DECLARE AI USE · KEEP HUMANS IN THE LOOP.

[Audio] Seven decisions before the study can move 1 Multi-year e-Dispatch data? 2 Precise incident locations? 3 Deployment + availability records? 4 Static, temporary, or dynamic positioning? 5 Authorized positioning sites? 6 Actual vs. estimated travel times? 7 Comparable predictive-prescriptive studies?.

[Audio] Evidence note Preliminary, transparent, and ready to be tested. REFERENCES STATUS Not yet supplied Claims + hypotheses are preliminary Ongoing work will prioritize traceable sources, explicit assumptions, and evidence-aware framing..

[Audio] Thank you for your attention. Questions, critique, or feedback? NEO · GROUP D.