About This Project

What This Is

Exceedance is a public dashboard that watches climate news. We monitor RSS feeds from more than a dozen outlets, Carbon Brief, Nature, NOAA, NASA, Reuters, and others, and pull in every article that touches on climate. Then we classify them by relevance, score their significance, group related stories into event clusters, and present everything on a single page.

We also maintain a curated set of climate indicators with status assessments. These indicators are tracked manually. The dashboard does not fetch raw scientific measurements. It shows you where the indicators stand and what the news says about them.

Every article links back to its original source. Every event cluster tells you which outlets reported on it. We think that matters. We think you should be able to trace a story back to where it started.

The project was inspired by the Project 2025 Tracker, but pointed in a different direction. Instead of tracking political promises, we track the climate story as it unfolds in the press. What scientists announce, what reporters find, what governments decide, what the numbers show when someone finally publishes them.

Dashboard Overview

The dashboard is built from several pieces that work together, each one showing you a different angle on the same story:

• Progress Bar — A single number that says how far along we are. Think of it as a thermometer for civilization. 0% would mean a safe, stable climate. 100% means impacts at a scale that reshapes how humans live. It's an inverse progress bar: progress here means things are getting worse.
• Indicator Grid — Climate indicators with individual cards. Global temperature anomaly, Arctic sea ice extent, CO₂ concentration, ocean heat content, sea level rise, coral bleaching, and more. You can filter by category to focus on what interests you.
• Event Map — A world map with colored dots marking where climate events are happening. Wildfires, floods, heat waves, policy announcements. Each dot links to the story behind it.
• Event Clusters — When multiple news outlets report on the same event, we bundle those stories together. One wildfire might generate twenty articles. This groups them so you see the event, not the noise.
• Media Feed — Individual articles, each tagged with significance, source credibility, and how fresh it is. You can filter by source quality to focus on the most reliable reporting.

Event Clusters

Here's something we noticed: when a major climate event happens, dozens of outlets write about it. The same flood, the same wildfire, the same record-breaking temperature. Reading twenty articles about the same event doesn't give you twenty times the understanding. It gives you the same understanding twenty times.

So we group them. A language model reads articles as they come in, and when it finds stories that are clearly about the same underlying event, it bundles them together into what we call a cluster. Each cluster gets a name, a category, a summary, and a count of how many independent outlets covered it.

Multi-source clusters (two or more outlets) appear as expandable cards at the top. Single-source stories show as compact rows below. Both are organized by time: today, this week, this month, and older. It's not perfect, but it's better than reading everything yourself.

Each cluster includes:

• Event Name — A human-readable title for what happened
• Category — Temperature, weather, ocean, ice, energy, policy, ecosystem, or disaster
• Source Count — How many independent outlets reported on it. More sources means more confidence.
• Summary — A brief description of the event, written by the pipeline
• Location — Where it happened, when we can determine it

Event Map

The map is a Leaflet canvas with dark tiles that match the dashboard's dark theme. In light mode, the tiles invert automatically. Events with known locations show as colored circles, clustered at low zoom levels so they don't overlap. Click a marker and you'll see the event name, category, source count, summary, and a link to the full cluster of articles.

Color tells you category: red for temperature, orange for weather, blue for ocean, light blue for ice, yellow for energy, purple for policy, green for ecosystem, lavender for science. Recency shows as opacity: recent events are bright, older ones fade.

A timeline slider lets you filter by date. Not every event has a geographic location. Some stories are about policy, or global trends, or scientific findings that don't map to a single place. Those stories still appear in the Event Clusters and Media Feed. They just don't get a dot on the map.

Media Feed

The media feed is where you see individual articles. Each one comes with several pieces of metadata that help you decide how much weight to give it:

• Source Credibility Dot — A small colored dot next to each source name. The color maps directly to the credibility hierarchy below. Gold means IPCC. Blue means NASA or NOAA.
• Title — Linked to the original source, so you can read the full article.
• Significance Badge — A colored label indicating importance: high, medium, or low.
• Freshness Indicator — Shows how recent the article is. Climate news doesn't go stale the way political news does, but it does age.
• Category Tags — Derived from which indicators the article relates to.
• Summary — A brief description of what the article is about.

You can filter by source credibility tier to focus on the most reliable sources. Sometimes you want to see everything. Sometimes you only want the gold-standard stuff.

Significance Levels

Not every story carries the same weight. A record-breaking heat wave that kills hundreds matters differently than a preliminary study about a new finding. Significance scoring helps you focus your attention. Scores are based on source credibility, urgency language, recency, and category diversity.

Source Credibility

Not all sources carry equal weight, and we think you should know that. Each article shows a small colored dot next to the source name. That dot maps to a credibility hierarchy that we maintain based on editorial standards, peer review, and institutional reputation. Here's what the colors mean:

Freshness Indicators

Climate news doesn't always go stale the way political news does. A story about a temperature record from last week is still worth reading. But eventually, articles do get old enough that they might not reflect the latest developments. We track freshness in two ways:

Dashboard freshness badge (in the navigation bar) tells you how recently the pipeline ran:

Article recency (shown on individual articles in the Media Feed) is based on when the article was published:

Status Scale

Each indicator is placed on a scale that tells you how close it is to a catastrophe milestone. The scale runs from "things are roughly where they've been" to "things have changed in ways that reshape civilization." Here are the levels:

Progress Bar

The progress bar is the paradox at the center of this dashboard. Most progress bars measure something getting better. This one tracks something getting worse. Zero percent would mean a safe, stable climate. One hundred percent means impacts at a scale that rewrites the rules. Every percentage point between here and there is a measure of how much the story has changed. We track it because ignoring the news doesn't make it go away.

Data Sources

The indicators on this dashboard draw from institutions that have been watching the planet for decades. The news articles we track come from many of the same places. These are the people who wake up every morning, check the data, and then tell the world what they found:

• IPCC AR5 / AR6 — The Assessment Reports. Thousands of scientists, one conclusion.
• NASA Climate — Vital signs of the planet, measured from space and ground stations.
• NOAA Climate.gov — Climate monitoring from the United States' environmental agency.
• Copernicus — Europe's Climate Change Service, part of the Copernicus programme.
• NSIDC — The National Snow and Ice Data Center. They count the ice so we don't have to.
• NOAA GML — The Global Monitoring Laboratory. CO₂ and methane, measured at the top of mountains.
• Global Carbon Budget — Annual emissions data, the accounting of what we put into the atmosphere.

Open Source

This project is open source because the data belongs to everyone. The code is MIT licensed, the data is CC BY-SA 4.0. If you want to build something with it, you're welcome to. If you want to improve it, even better.

Pull requests are welcome. Every contribution makes the picture a little clearer. And if you just want to read through the articles and think about what they mean, that's welcome too. Not everyone has to build something. Sometimes the most important thing you can do is pay attention.