The evidence layer for humanitarian decision-making.

Where most humanitarian dashboards present a slice — funding alone, displacement alone, one agency's view of the operation — PRISM joins them. INFORM severity sits next to FTS coverage; ACLED conflict events sit next to IOM DTM displacement totals; the GHO appeal sits next to the actual ReliefWeb reports that justify it. Decisions stop happening in isolation; trade-offs become visible. The current stage addresses an integration gap and a missing evidence base for human rights monitoring.

The platform synthesises data from OCHA (FTS, HPC.tools, CERF, CBPF), ACAPS, ACLED, UNHCR, World Bank, UNICEF, IOM DTM, IDMC, IPC, INFORM (JRC DRMKC), GPC, OECD DAC, HDX HAPI, EM-DAT, GDACS, NASA FIRMS, ReliefWeb, Copernicus ERA5, ASAP, Eurostat, the IOM Missing Migrants Project and the EU migration-funding databases. Live feeds refresh in cycles from 15 minutes to a month: PRISM's daily refresh sweep refetches any feed whose copy is more than 24 hours old, and the committed FTS donor snapshot is rebuilt monthly. The migration-funding databases are loaded from their publishers' releases. The analysts, authors and editors who file alongside the feeds are listed on the contributors page.

Humanitarian data infrastructures are rarely examined as sites of epistemic politics. Yet the categories through which need is assessed — who counts as 'refugee' versus 'migrant', what constitutes 'protection risk', which actors produce 'legitimate' knowledge — are not neutral inputs into resource allocation decisions. They are products of unequal epistemic arrangements that systematically marginalise crisis-affected populations and local responders while embedding institutional hierarchies into algorithmic form.

This is measurable, and the system's own reporting measures it. In 2025, as global humanitarian funding collapsed, severely underfunded crises saw only 49% of their needs prioritised, against 64% for moderately underfunded ones — the response deepened the inequity it was meant to correct (OCHA, GHO 2025). Nearly a decade after the Grand Bargain promised localisation, local actors held 11% of Humanitarian Country Team seats and 7% of Inter-Cluster Coordination Group seats (Grand Bargain self-reporting, 2024). Funding follows the history of intervention more faithfully than it follows current need. Who is allowed to produce knowledge decides who receives resources — and the people closest to a crisis are asked to supply 'beneficiary feedback' to frameworks designed elsewhere, while their credibility is questioned because of where it was earned.

PRISM in its current form is not, and has never been, the end objective. The current stage addresses the integration gap. The next stage — to collect its own data, to democratise knowledge, to amplify the voices of local actors in a space ruled by the cruel nature of political actors — has begun. The contributor desk is open: protection incidents filed by people who know these crises firsthand, held to one standard — show how you know — reviewed for evidence, and preserved in a record that accumulates and never deletes. The platform that began by serving institutional datasets is learning to make its own.

It plugs into the data-collection tools field teams already use, anonymises records on the way in, encrypts them end to end, and triangulates the findings against PRISM's wider humanitarian and development data. The result is usable evidence the same day data lands: real-time data-quality monitoring and enumerator KPIs for large-scale surveys, a single dashboard for cross-country prioritisation, and donor-proposal intelligence trained on 25 institutional donors' rules.

Each project is private to its assigned team and never appears publicly. Dashboards show aggregate findings only, with no personal data.