About Mind the Gap
Mind the Gap is an AI-powered information platform that aims to make global inequality, humanitarian crises and development issues easier to understand and turn awareness into action.
Mission alignment
The project's primary focus is UN Sustainable Development Goal 10 (Reduced Inequalities), with SDG 1 (No Poverty) and SDG 3 (Good Health and Well-being) as secondary goals. It is inspired by the values and mission of Oxfam.
How summaries are produced
News summaries are written to be understandable at a glance and always link back to their original source. As the project develops, summaries will be generated with AI assistance from underlying source material — every summary will continue to disclose and link its source, and will never state something the source doesn't support.
Where the map data comes from
Each country page shows two different kinds of numbers, and they're kept separate on purpose.
Structural indicators — population, the Gini index, the poverty headcount ratio, undernourishment — are fetched live from the World Bank Open Data API by pipeline/fetch_worldbank_data.py, re-runnable any time to pull the latest published figures. Where a country has a national statistic worth citing (for example, Nigeria's National Bureau of Statistics Multidimensional Poverty Index), that's included too, with its own source link.
The map's colour (need level) is not based on those structural indicators. That's a deliberate choice: for countries in active conflict, like Sudan and Yemen, the last time a normal household survey was possible was years before the current crisis — so the "official" poverty rate looks lower than it's ever been, right when the actual need is at its highest. Using that number to colour the map would show the world's worst crises as low-need. Instead, the need level is driven by the most recent, specific, cited estimate of how many people are directly affected by the crisis described on that page — shown at the top of the "Current crisis" box on each country's page, with its source and date.
We tried to pull that crisis figure live too, from ReliefWeb's and HDX's public APIs. Both blocked every automated request made while building this pipeline (HTTP 403, even for a plain, documented query) — the fetch code is still in the pipeline script, disabled, in case that changes. Until then, those figures are entered by hand each time the pipeline is refreshed, always with a citation and a checked date, which you can verify by following the source link.
The exact calculation: the cited crisis figure divided by the country's total population, placed into one of five bands (35%+ of the population → level 5, 20–35% → level 4, 10–20% → level 3, 4–10% → level 2, under 4% → level 1). This is a real, deterministic formula — not an editorial judgment call — and it's implemented in pipeline/fetch_worldbank_data.py, which anyone can read and re-run. If a country has no cited crisis estimate, the pipeline leaves the figure unset rather than defaulting to a numeral, and the site shows "Insufficient data" instead of a misleadingly specific number.
That gap — between what official statistics can currently measure and what's actually happening on the ground — isn't a bug we're hiding. It's part of what this project is trying to show.
Known limitations we're actively working on
Being transparent about the method means being transparent about where it's still imperfect:
- Mixed measures across countries. The "affected" figure means something different per country — food-insecure in Sudan, internally displaced in Haiti — but the formula divides all of them by population the same way. That inconsistency is visible (each country's "measured" text says exactly what its number counts) but not yet normalized.
- Regional crises divided by national population. When a crisis is concentrated in part of a country (like northern Nigeria), dividing by the whole country's population can understate it relative to a country where the same number of people affected is a much larger share nationally.
- No population figure for Gaza on its own. World Bank Open Data only publishes population for "West Bank and Gaza" as one combined territory, not Gaza separately. Any Gaza-specific crisis estimate divided by that combined figure would be diluted and misleadingly low, so Gaza's profile deliberately leaves the need score unset rather than publish a distorted number.
Found something wrong, or have feedback?
Tell us. Submitting sends an email directly to go93mof@mytum.de — a real inbox someone on the team reads, though we're not yet committing to a specific response-time target. Limited to 3 messages per 24 hours per device, to keep this from being spammed.
Photo credits
Country photos are real photographs from Pexels, chosen as dignified, representative images of each place rather than staged "crisis" stock photography. None depict the specific event described in that country's news — they're illustrative, not documentary. Free to use under the Pexels License (attribution not required), credited here anyway:
A university project
Mind the Gap is built as a university project exploring how technology can make complex global issues more accessible.