Leith Research — climate and energy analysis, built from the data up

Climate and energy analysis,
built from the data up.

We do three things. Machine learning that estimates the energy efficiency of every building in a country. Transition-risk models that connect climate scenarios to real power plants and companies. Trade analysis that shows where a country's clean industries can grow. Our methods and code are open source by default, so you can check the work and keep running it without us.

What we do

Machine learning for building efficiency

Most buildings have never had an energy assessment. We train models on the ones that have, then estimate energy use and efficiency for every building in a country, with honest uncertainty. Governments use the results to target retrofit grants, design energy bills and levies, and cost national policies.

Climate scenarios, assets and transition risk

Climate scenarios describe whole economies. A bank or a supervisor needs to know what they mean for one power plant, one factory, one company. We build the data and models that make that link: which assets earn less as the energy system changes, whose revenues are exposed, and by how much under each scenario.

Trade and industrial-policy analysis

Which clean technologies can a country realistically make and export next? We answer that from world-trade data on every traded product, so ministries can back industries that build on what the country already does well, rather than guess.

Selected work

Cards marked "View chart" open one figure from the work, aggregated so it can be shared publicly. Cards marked "Code" link to the public repository.

3.9Mbuildings

An energy-efficiency estimate for every building in a country

Fewer than one in ten Czech buildings holds an energy performance certificate. We linked the 262,000 that exist to administrative registers and metered consumption, then trained a model that estimates energy use and an efficiency class for all 3.9 million buildings, with a probability for each class rather than a single guess. Delivered to the Czech Ministry of Finance as an interactive map down to municipality level. We also checked around 1,700 public buildings' certificates against their metered bills to see how far the paperwork sits from reality.

machine learningrecord linkageCzech Ministry of Finance

View map →

156electricity market zones

What a power plant actually earns under a climate scenario

Climate scenarios give one average electricity price per country, but a solar farm doesn't earn it and a gas plant earns more. We build the data and models that Theia Finance Labs, a non-profit that runs climate stress tests for financial institutions, uses to work out what each plant would actually earn under each scenario, for power plants worldwide.

electricity marketsclimate scenariosTheia Finance Labs

View chart →

516green products assessed

Where Czech clean-tech exports can grow

An interactive map of the Czech Republic's export potential in green technologies, built from world-trade data on every traded product and an economic-complexity model of what the country can plausibly make next. Used by 2ET, a Czech economic-policy think tank, to decide which industries to support.

world trade dataeconomic complexityopen source2ET

View chart → Code →

331local authorities

Who pays when energy levies move from electricity to gas

Estimating gas and electricity use by house type, heating system and local authority across England and Wales, from census heating data, floor-space statistics and metered consumption. An input to the debate on rebalancing policy costs between electricity and gas bills.

censusmetered consumptionopen sourceUK

View chart → Code →

Transition exposure of small companies, for the banks that lend to them

tilt helps banks understand the climate transition exposure of the small and medium-sized companies they lend to, by mapping what those companies make to emissions data and transition indicators. We supported the data engineering behind it.

SME transition riskdata pipelines

tiltsmes.org →

Open by default

We publish our methods, code and data notes unless a client's data prevents it, and we hand over work the client can keep running without us. No licences, no black boxes.

MapaPrilezitosti — the Czech green-export opportunity map: trade-data pipeline, complexity model and the app.

LevyRebalancingGIS — energy use by house type, heating system and local authority for England and Wales.

LeithResearchPage — this site, including the aggregated data behind every chart.

Team

Leith Research was set up in Edinburgh in 2024. Leith is Maxim Oweyssi and a small circle of collaborators brought in per project.

Maxim Oweyssi

Founder

Physicist by training: at the Max Planck Institute for Astronomy he estimated the masses of galaxy clusters from X-ray images. Since moving into climate and energy policy he has built building-stock models for the Czech Ministry of Finance, energy-bill analysis in the UK, trade-opportunity maps for Czech industrial policy, and asset-level climate-risk data with Theia Finance Labs.

Jakub Červenka

Collaborator, climate stress testing

Economist working on climate and financial risk. He leads the 1in1000 climate stress-testing programme at Theia Finance Labs in Berlin, run with the University of Oxford Sustainable Finance Group, and advises central banks and development banks on climate stress tests. Previously stress-test modelling at the European Central Bank and risk advisory at Deloitte. With Leith he works on transition-risk and scenario projects.

Climate and energy analysis,
built from the data up.