Method and limits
Minbar Watch records what states tell mosques to preach on Fridays. Many governments publish a unified sermon, الخطبة الموحدة, every week. Others set only the topic, or send occasional directives. We collect those titles from official pages, tag them, and line the countries up by week.
What is collected
For each entry: the Friday it was preached, the original title exactly as published, an English translation, a one-sentence summary, the type (sermon or directive) and a link to the page it came from. Entries taken from a news report rather than the authority's own site are marked secondary. Dates we could not confirm are marked low confidence.
The four systems
Topics and themes
Each entry gets one of eight topics, which sets its tile colour, and up to three finer themes. Tagging is done by keyword rules over the original title and its translation, then corrected by hand where a rule got it wrong. Hand-corrected entries are marked. The rules are in pipeline/taxonomy.py.
Convergence
Two countries converge when they preach the same theme on the same Friday. Themes are narrow (for example "Drugs & addiction" or "National days"), so a match is a real signal. Three catch-all themes (ritual and occasions, belief and piety, general good manners) are too common to count and are left out of the matching. Sharing a broad topic is weaker evidence: most sermons everywhere are about faith and worship, so the pairs table reports lift over chance alongside raw agreement.
Limits
- Coverage is uneven. An empty tile means nothing was collected, not that nothing was preached.
- A title is a thin signal. The same title can carry very different sermons, and political content often sits in the second half of a sermon on a devotional subject.
- Religious calendars create shared themes with no coordination behind them: every country preaches on Ramadan in Ramadan.
- Shared themes show correlation. They do not show that one government copied or instructed another.
Data
The full dataset is one file: data/data.json.