Celebrating 100 Topics (and richer, more accessible search)
100 Topics. 34 partners. Hundreds of hours of research saved. Here is what it adds up to.
Every day, researchers, policymakers, and analysts try to make sense of how the world is responding to climate change. They search across thousands of laws, policies, and financial instruments - and too often come up short, not because the information doesn't exist, but because it is buried in inconsistent language, synonyms, and acronyms. Climate Policy Radar's Topics solve this by packaging expert knowledge - related terms, regional variations, and contextual nuance - so a single search surfaces everything relevant. We've now built 100 of them, giving anyone who uses our database something close to a complete map of how the world is responding to climate change.
Beyond keywords: building a shared language for climate search
The global landscape and language of climate change is complicated. Themes and terms are deeply interconnected, and the same concepts are expressed in many different ways depending on who is speaking, where they are located, and who they’re communicating with. Compounding this, climate documents are often dense, technical, and acronym-heavy. For anyone trying to understand how the world is responding to climate change, finding what actually matters across these documents is harder than it should be - resulting in missing critical information, or spending significant time manually working through related terms.
Search for “climate adaptation” in a set of documents and you might miss references to “increasing climate resilience” or “reducing climate change impacts”. Search for “youth” and you may not surface mentions of “children”, or “future generations”, or “girl” or “boy”, therefore missing whole swathes of policy that mention young people in different ways.
Our Topics function is how we address this. When you search the Climate Policy Radar database using a Topic, rather than returning results based on a keyword match alone, that Topic encodes expert knowledge containing related terms, regional variations, and contextual nuance across thousands of documents.
Topics provide a much richer and more useful search, at speed and scale. They make it possible to search like an expert, without necessarily needing to be one. We sometimes describe them as “smart search terms”, one word or phrase connected to a whole web of meaning, assembled through in-depth research, critical questioning, and extensive collaboration.
"Understanding developments in climate policy around the world involves looking through a vast and heterogeneous array of documents, which often feels like searching for a needle in a haystack. CPR’s Topics can help us navigate this challenge more effectively, so that we spend more time analysing policy responses (instead of searching for needles!). As the climate policy landscape continues to evolve rapidly, these tools will be increasingly valuable for our work.”
Emily Bradeen, Policy Officer, Grantham Research Institute on Climate Change and the Environment, LSE
What Topics unlock and how they’re being used
Our database covers most of the major Topics across climate law, policy, finance, and governance - from targets and policy instruments, to different economic sectors, to finance flows, to renewable energy.
Each Topic works as a rich search term. However, we know that climate policy is rarely about a single Topic in isolation - and our search functionality reflects this. We have intentionally classified at the paragraph level in all our documents, which means Topics can be layered together to build increasingly precise searches.
For example, if you are interested in which countries have implemented specific minimum standards within their policies to support the renewable energy transition, you can combine "renewable energy" and "codes and standards", then apply the "policies" document type filter to move from millions of passages to a highly specific set of results in a few clicks. This is one of the ways in which our database and tools are genuinely useful for complex research questions, alongside broader discovery.
So how are Topics being put into practice?
The Convention on Biological Diversity (CBD) is using CPR to analyse country submissions and assess their progress towards the 23 targets in the Kunming-Montreal Global Biodiversity Framework - a monumental task (some reports are 1000+ pages). The analysis will be presented in Nairobi in August, feeding into the global review of progress towards 2030 targets to be negotiated at CBD COP17.
The Global Green Growth Institute (GGGI) recently used CPR to analyse how well G20 countries had incorporated just transition components into their Long-Term Low Emission Development Strategies (LT-LEDS). Their team applied our ”just transition” Topic to cross-check findings at scale, work that also helped train and refine the Topic itself. The results fed into this report, which identifies strengths, gaps, and good practice to help countries develop fairer and more inclusive long-term climate strategies.
The Grantham Research Institute at LSE uses CPR across its climate law and policy research. This includes supporting their annual Global Trends in Climate Change Litigation report, one of the most influential publications in the climate law space. Our tools are helping the team to analyse a larger set of cases more accurately, test expert hypotheses, and visualise trends and comparisons over time, adding to the depth and richness of their research. The 2026 report will be launched at London Climate Action Week this June.
Researchers working on sustainability transition policy at the Friedrich-Alexander University (FAU) used CPR to gather data on climate and renewable energy targets for the Climate Policy Atlas, an open access, machine-readable dataset visualising climate and energy policy analysis across the EU and UK. CPR’s breadth of historical targets allowed the team to track how each country’s ambitions have changed over time, and explore whether more ambitious targets translate into better outcomes. This work also feeds into Germany's national research data infrastructure for energy, NFDI4Energy, supporting energy research and evidence-based policymaking.
Research, consult, train, refine: how Topics are built
The reason our users and partners trust our Topics comes down to how they're built. Building a Topic is not a matter of simply feeding terms into an algorithm. This is a rigorous, iterative, and human-first process, built collaboratively with users to ensure that topics actually represent the community’s specific needs.
It starts with internal research: drawing on existing taxonomies from across the climate and environmental landscape, reviewing how organisations, academics, and civil society groups define and discuss a Topic, as well as any recent developments in this area. We deliberately seek out a representative range of perspectives, including voices that challenge dominant framings, so we can understand where they fall short and how we can do better.
Everything is deepened and refined through ongoing collaboration with external experts. Definitions, sub-Topics, and connected terms are developed together, groundtruthed and stress-tested, until we reach a consensus on what to include and why. The 34 organisations who have contributed to our Topics so far are not just informing CPR’s work; they are shaping how climate concepts are defined and used across the fields they work in.
“In a fast-evolving and expansive field like climate action, bringing in external expertise helps ground CPR’s tools in the best available knowledge while keeping them highly relevant for researchers, investors, policymakers, and others navigating the global policy landscape."
Jose Diaz, Data Analyst, Climate Policy Initiative
Only after that does AI technology come in. A classifier model is trained to find relevant passages across the documents in our database, reviewed and tested by our team who manually label a sample of documents first to make sure it’s performing well and picking up what it needs to for our users. We intentionally choose the least energy-intensive models that meet our needs, which in practice means most models rely on efficient keyword-based detection.
We store all Topics (current, and future) in our public concept store, providing a transparent, auditable record of the full network and terms that underpin each one. This also offers the possibility for others to build their own.
What’s next: more Topics, more collaboration, and a shared digital infrastructure
We continue to build. The next major Topics in development are “mitigation”, covering the direct interventions and enabling conditions that reduce emissions and enhance sinks of greenhouse gases, and “climate justice”, identifying arguments of how climate change impacts people, communities, and countries differently and disproportionately. This is a decent set of building blocks for understanding climate policy - and we know we have much more to do.
We are also developing more granular Topics to allow users to analyse laws and policies in more detail. For example, we have deepened our ‘risk’ Topics so that users can not only search for “extreme weather event” but also for types of extreme weather such as “drought” and “heavy precipitation”.
Many of our current Topics already serve users working across climate and nature - including “subsidies”, “finance flows”, and “indigenous people”. This year, we are deepening that support by developing nature-specific Topics, starting with “ecosystems” to help users find information on ecosystem types like mangroves or grasslands, followed by “nature-based solutions” - bringing together actions to address climate change and biodiversity, and breaking down the siloes between these interconnected domains.
Partners like GIZ, NDC Partnership and the International Union for Conservation of Nature (IUCN) are already using CPR to assess how coherent commitments are across climate and nature, and our nature Topics will advance this work. We are also progressing our analysis and data visualisation capabilities, so that more users can explore trends and patterns across the database directly.
Longer term, one of our priorities is finding smarter ways to maintain the rigorous approach and expert validation that makes our Topics trustworthy, but at a greater scale. This includes exploring how large language models can handle routine classification work under supervision, freeing our teams to go deeper on the concepts that require the most nuance.
But the more important point is this: our pipeline and methodology are openly available. Any organisation can take what we have built and train their own classifiers, on their own terms. The ONE Campaign is doing exactly that, building a health finance taxonomy in our concept store to explore how their data could link to policy context. That is what it looks like for CPR to function as digital public infrastructure, strengthening the wider ecosystem’s ability to generate insight and build solutions on top of shared, structured evidence.
Explore our Topics at app.climatepolicyradar.org/search, and get in touch with our team to discuss how they could support your work - or if you would like to be involved in future Topics development.
