Introduction and Research Context
In complex humanitarian emergencies, the static deployment of resources often fails to address the rapidly shifting vulnerabilities of affected populations. My research project focuses on developing an algorithmic Explosive Ordnance Risk Education (EORE) Prioritization System for the Occupied Palestinian Territory (oPt). Currently, EORE interventions lack a dynamic mechanism to align with emergent, localized needs. By synthesizing multi-source humanitarian indicators into a unified spatial model, this project aims to transition UNMAS operations from reactive tasking to proactive, data-driven resource allocation. This post outlines the methodologies employed in system development and the formal framework designed to assess and evaluate its operational outcomes.
System Architecture and Automated Data Pipelines
To ensure the resulting system is interoperable, cost-effective, and highly scalable, I utilized a robust geospatial technology stack integrated directly with the UNMAS Information Management System (IMS):
- Geospatial Database Architecture: The system is built upon ArcGIS Enterprise and PostgreSQL. This strategic architectural choice eliminates extraneous software costs and ensures seamless integration with established territorial and neighbourhood gazetteer data.
- Automated API Integration: I engineered custom Google Apps Script modules to interface with the ArcGIS REST API. This automates the fetching, spatial processing, and bidirectional database updates in controlled batches, ensuring API stability and reducing manual processing overhead.
- Algorithmic Spatial Processing: A critical technical challenge was dynamically mapping coordinate-based incident and population data to specific neighbourhood polygons. To solve this, the pipeline utilizes Bounding Box pre-filtering combined with a custom Ray-Casting algorithm. This allows for highly optimized Point-in-Polygon (PIP) spatial joins, accurately aggregating micro-level data (such as ACLED events and EHA requests) directly into neighbourhood boundaries.
Spatial Multi-Criteria Decision Analysis (SMCDA) Matrix
The analytical core of the system is an 8-variable scoring matrix that processes real-time data to quantify neighbourhood vulnerability. The system dynamically evaluates the following indicators to generate a comprehensive risk profile:
Indicator | Data Source | Scoring Logic (Max Score) |
|---|
1. EO Confirmed | UNMAS IMS | EOD confirmed presence = 5 |
2. OCHA Population | OCHA | > 85,000 = 5; > 42,000 = 4; 1-42,000 = 3 |
3. Displaced Population | Site Management | > 5,000 = 5; 501-5,000 = 4; 1-500 = 3 |
4. EHA Conducted Risk | UNMAS IMS | High Risk = 5; Mid Risk = 3; Low Risk = 1 |
5. ACLED EO Events | ACLED | > 0 events mapped via PIP = 4 |
6. EO Incident w/ Victims | UNMAS IMS | Incidents post-Oct 6, 2023 = 3 |
7. EOD Recommends EORE | UNMAS IMS | Recommendation logged in EHA = 3 |
8. EHA Requests | UNMAS IMS | 100-day rolling window: > 3 requests = 2; 1-3 = 1 |
The algorithm calculates a total_scor for each neighbourhood. Areas are then automatically classified into actionable intervention tiers: Very High (20+), High (15-19), Mid (10-14), and Low (<10).
Planned Approach to Assessing and Evaluating Project Outcomes
To rigorously evaluate the success of this research project, I have designed a triangulated evaluation framework measuring technical efficacy, stakeholder acceptance, and operational impact:
- Empirical Technical Validation: The system's algorithmic integrity will be assessed through continuous monitoring of API payloads and execution logs within Google Apps Script. Evaluation criteria include the zero-fault execution of the Point-in-Polygon ray-casting algorithm, successful batch updates, and the accurate dynamic recalculation of the 100-day rolling windows without data latency.
- Qualitative Heuristic Evaluation (Stakeholder Adoption): In project management, a tool's viability relies heavily on end-user adoption. I will evaluate the system's operational relevance via structured feedback cycles with the OPT EORE Technical Working Group and the Mine Action AoR. By deploying automated email intelligence reports (sorted multi-level by Governorate, Classification, and Recency of Activity), I will assess whether the programmatic outputs validate their on-the-ground intelligence and meet operational requirements.
- Longitudinal Impact Assessment: The ultimate metric of success is a quantifiable shift in humanitarian intervention. I will evaluate this outcome utilizing the ArcGIS Web Application developed alongside the algorithm. By monitoring the tasking behaviors of Implementing Partners within the application, I can quantitatively track whether actual EORE deployments have demonstrably increased within the identified "Very High" priority polygons. A measurable realignment of field resources to match the system's data will signify complete project success.
References
- ACLED (2024) Armed Conflict Location & Event Data Project. Available at: https://acleddata.com (Accessed: 25 October 2024).
- OCHA (2024) Occupied Palestinian Territory Population Data. Available at: https://www.ochaopt.org (Accessed: 25 October 2024).
- Project Management Institute (2021) A Guide to the Project Management Body of Knowledge (PMBOK Guide). 7th edn. Newtown Square, PA: PMI.
- UNMAS (2024) Information Management System (IMS) Spatial Data. Internal UNMAS Database.
