MoodAtlas

Inspiration

Depicting the Mood of the World Through Google Search Trends Every day, billions of people turn to Google not just for information, but for something far more intimate — to make sense of what they are feeling. Before confiding in a friend, before naming an emotion out loud, people type it into a search bar. That quiet, private act of searching is, in aggregate, one of the most honest records of human emotional life we have ever had access to.

This project started from a simple but profound question: what if we could take the emotional pulse of the world? Not through surveys, which people answer carefully and publicly, but through search data — the unfiltered, uncurated language people use when they think no one is watching.

Google Trends does not tell us what people feel with clinical precision. It tells us what people are reaching for — which emotions they are trying to understand, name, or escape. A spike in searches for "feeling overwhelmed" is not just a data point; it is millions of people simultaneously arriving at the edge of their capacity. A surge in "feeling grateful" tells a story of collective relief, or perhaps of a cultural moment prompting people to count what they still have.

What it does

By mapping these searches across time and geography, patterns emerge that no single person could see alone. Anxiety clusters around uncertainty — elections, economic shifts, pandemics. Sadness deepens in winter months. Loneliness peaks, paradoxically, during the most social times of year.

What we learned

What this means is both humbling and hopeful. It means that the emotions we experience as deeply personal are, in truth, deeply shared. The person searching "feeling lonely" at 2am is not alone in that search. And in understanding the collective mood of the world, we take the first step toward something important: compassion at scale — the recognition that beneath every search query is a human being trying to feel a little less lost.

How we built it

moodAtlas was built by harnessing the raw emotional signal hidden inside billions of Google search queries, transforming everyday searches into a living map of how the world feels. Powered by Zerve AI agents, it interprets that data in real time — turning numbers into narrative, and trends into a window on the human condition.

Challenges we ran into

Wrangling Google Trends' strict API rate limits and narrow query windows meant we had to architect clever batching and caching strategies to extract emotionally meaningful signal without hitting dead ends.

Accomplishments that we're proud of

We turned something as fleeting and invisible as a search query into a coherent, real-time portrait of global human emotion — and made it beautiful enough to actually feel something when you look at it.

What's next for MoodAtlas

We can expand MoodAtlas into live mood tracking across languages and cultures, building predictive models that can anticipate emotional shifts before they peak — moving MoodAtlas from a mirror of how the world feels to an early warning system for when it needs help.

Built With

• python
• zervehack