Every day, for every country, we look at what people read, search, watch, listen to, play and buy online, and find what stands out. Here's how, in plain words.
What we measure
We use public signals, published for each country, of what people:
- Read: the articles people look up most.
- Search for: what's trending in searches.
- Watch: what's topping streaming charts.
- Listen to and download: music and app charts.
- Play and buy: best-selling games.
- Expect: what prediction markets are trading on, shown as market odds.
Each source is credited next to the data it provides.
These show attention, not opinion. Reading about something doesn't mean liking it, and online behaviour isn't a poll of a whole country.
One topic, many languages and sources
People in Brazil read about a series in Portuguese, people in Japan search for it in Japanese, and it also sits in a streaming chart. We recognise that these are the same topic, so you see it once, with every place it shows up.
Spotting what's new
Big topics are always big: a country's most famous football club or its president is read about every day. So we don't just rank by volume. We compare each topic with its own normal level in that country, and label the ones that stand out:
- New: it wasn't on anyone's radar until now.
- Breakout: a sudden jump far above its usual level.
- Rising: steadily climbing for several days.
- Back: returning to a chart it recently left.
- Falling: clearly fading.
- Weekly / Yearly: a repeat, like a weekly TV show or an annual holiday. These are shown but not counted as news.
Because every country is compared with itself, small countries aren't drowned out by big ones.
Ranking a country's topics
Each country's list combines how big a topic is with how fast it's changing. A topic that shows up in several kinds of sources (say, both searches and charts) counts for more than one seen in a single place. We put every topic in one of 20 categories, from politics to gaming, so you can see what's happening in each part of life.
Confidence and limited data
Some signals are stronger than others. A topic seen in only one source, with no history to compare against yet, is marked Low confidence. Countries where few sources are available are marked Limited data, and we say so rather than guess.
Where topics are heading next
Topics often travel: what rises in one country often reaches its neighbours or countries with shared language or culture a few days later. From our daily archive, we learn how topics have spread between countries in the past and estimate where a rising topic is likely to go next. These are estimates, not promises. Simultaneous worldwide releases are not counted as spreading.
AI briefs
Each country's short brief is written by an AI model. To keep it honest:
- It sees only that country's data on the page, nothing else.
- Every sentence must point to the topics it's based on.
- It may not invent reasons. It can only say why something is trending when a linked news headline says so.
- It may not copy headlines, mention things from other countries' data, or use loaded language.
- Every brief is checked automatically before it's published. A brief that fails gets no place on the page: no brief is better than a wrong one.
Briefs are always labelled AI-generated. Rankings, labels and charts are calculated, not written by AI.
What we can't see
- Each platform reaches some people more than others: phones, apps and game stores differ a lot from country to country.
- A few countries block or don't use these platforms, so their picture is thin or missing.
- Very small countries have less data, which makes their signals noisier.
The same method, every day
The archive is only useful if every day is measured the same way. When we improve the method, we note it in the corrections log.
Questions about the data? data@currentzeitgeist.com