MODERATE delivers an open, privacy-preserving marketplace that transforms fragmented building energy data into interoperable, actionable intelligence, putting analytics tools directly in the hands of facility managers, ESCOs, and policymakers.
Dr Cristian Pozza, Senior Researcher at Eurac Research | LinkedIn profile
Sofia Bazzano, Head of EU projects atREHVA | LinkedIn profile
(Note: Opinions in the articles are of the authors only and do not necessarily reflect the opinion of the European Union)
Buildings account for approximately 40% of energy consumption and 36% of greenhouse gas emissions in the EU, yet most operational building data remains locked in incompatible, proprietary silos. The MODERATE project tackles this challenge through a three-layer open platform architecture combining federated Data Nodes, machine-learning-based Data Synthetisation, and a suite of 11+ open-source Analytics Services. MODERATE has developed and validated tools spanning fault detection, solar potential assessment, energy benchmarking, local energy community siting, and measurement and verification of energy savings. A key innovation is the use of synthetic data generation to produce publicly shareable datasets that preserve the statistical properties of confidential real-world data, breaking the privacy barrier to open science. Deployed on datasets covering thousands of buildings across Germany, Italy, France, and Spain, the platform is reaching TRL 8 with all tools publicly accessible under permissive open-source licensing. This article presents the platform architecture, the analytical tools, real-world validation results, and the post-project sustainability model.
Europe's building stock is its most energy-intensive asset class, responsible for roughly 40% of final energy consumption and 36% of CO₂ emissions across the EU. Mandatory renovation targets under the recast Energy Performance of Buildings Directive (EPBD), the Energy Efficiency Directive, and the Fit for 55 package require data-driven decision-making at scale—from individual retrofit investments to district-level renovation roadmaps and the expansion of local energy communities (LECs).
Yet a fundamental barrier persists: building performance data is systematically fragmented. Energy management platforms are typically proprietary and vendor-locked; datasets collected by utilities, facility managers, and real estate operators cannot be combined; and the General Data Protection Regulation (GDPR) imposes strict constraints on sharing data that may contain personally identifiable information (PII). The result is that high-value operational intelligence sits unused in organisational silos, unable to inform benchmarking, fault diagnosis, or policy.
Three specific gaps drive this problem. First, there is no standardised, GDPR-compliant protocol for anonymising and sharing building performance datasets while preserving the statistical properties that make them analytically useful. Second, data owners lack economic incentives to open their data, because doing so transfers value to competitors without return. Third, almost all existing platform results from EU-funded projects are locked inside consortium IP agreements, preventing reuse.
MODERATE was designed to close precisely these three gaps: technically, legally, and economically.
Table 1. Key facts and figures of the MODERATE project. Source: MODERATE project.
MODERATE's platform is built around a modular three-layer architecture that separates data custody from data use, enabling a distributed ecosystem where data ownership is respected while insights are freely shared.
Data nodes are the entry points to the MODERATE ecosystem. Each node is deployed and operated by the data producer—a building operator, utility, facility manager, or public body—on their own infrastructure. This keeps sensitive raw data under the custody of its owner at all times. Nodes expose standardised APIs and adopt common data models evaluated during the project, including Brick Schema, building on lessons from H2020 predecessors (BuiltHub, BIGG, Matrycs). The IOTA Tangle distributed ledger provides an immutable audit trail of data usage licences from producer to consumer, enabling enforcement of access policies through smart contracts and building a trust layer for the marketplace.
This is MODERATE's most novel technical contribution. Machine learning models—primarily generative approaches—are trained on real building datasets and used to produce synthetic counterparts that preserve statistical distributions, temporal autocorrelations, and cross-variable dependencies, while containing zero personally identifiable information. The synthetic datasets can then be published openly on the marketplace while limiting re-identification risk.
This creates a two-tier data economy: a synthetic open tier accessible to all researchers and developers, and a premium tier of real, validated datasets available for commercial purchase. The synthetic data serves as both a GDPR-safe preview and a training ground for AI models; once developers have built and validated tools on synthetic data, those tools are directly compatible with the premium real datasets—closing a positive value loop that incentivises data owners to participate.
The technique, well established in finance and healthcare, had not previously been applied systematically to building energy performance data. MODERATE's contribution is a validated methodology and toolset specifically tailored to timeseries of energy consumption, indoor environmental quality, and lifecycle data.
The Analytics Layer comprises the open-source tools that convert data into actionable insights. All tools are released under permissive open-source licensing, ensuring commercial-friendly reuse with no copyleft constraints. Integration is handled through the MODERATE marketplace, which also manages discovery, access control, and usage accounting for both data and services.
MODERATE has developed and validated eleven analytics services, each targeting a specific professional need. The following describes the tools of greatest relevance to BUILD UP’s February Topic of the Month on data gap.
Developed in collaboration with project partners, this tool applies statistical decision-making and machine learning to timeseries data from Building Automation and Control Systems (BACS). It detects anomalous patterns in near real-time, enabling predictive maintenance before failures occur. Energy managers receive automated fault alerts with explanatory diagnostics via a web interface. In pilot deployments on commercial buildings, the tool identified heating, ventilation and air conditioning (HVAC) faults responsible for measurable excess energy consumption that had gone undetected for months.
A complementary tool to fault detection, the Contextual Anomaly Detector operates on energy consumption timeseries using the distance-based contextual matrix profile algorithm, combined with supervised and unsupervised approaches. Its self-tuning capability means it requires no manual threshold setting—particularly valuable for facility managers without data science expertise. The tool identifies abnormal energy usage patterns that are contextually unusual (e.g., anomalous consumption on a public holiday), going beyond simple threshold violations.
Using hourly electricity consumption data, this tool compares a target building against a peer group of buildings with similar consumption characteristics, selected algorithmically from the platform's dataset. The output is a ranked performance assessment with actionable KPIs, enabling facility managers to quantify the gap between current and best-practice performance and to prioritise interventions. The approach goes beyond static –EPC-based comparisons by using real operational data normalised across multiple years and climatic conditions, capturing the performance gap between design and in-use energy use.
Implementing Option C of the International Performance Measurement and Verification Protocol (IPMVP), this tool provides a standardised, auditable methodology for quantifying energy savings from efficiency measures. It is directly applicable to Energy Performance Contracting (EPC) and ESCO business models, where independent verification of savings is a contractual requirement. The tool reduces the cost and complexity of M&V, lowering a key transaction-cost barrier to EPC adoption.
Identifying viable sites for Local Energy Communities is a critical bottleneck in the energy transition. This tool uses geospatial analysis, load profile data, and renewable energy generation potential to screen building stocks and identify locations where LECs can be economically self-sustaining and generate social co-benefits. It was validated using datasets in the Valencian Community
An interactive map-based service enabling users to calculate the solar energy potential of individual buildings, evaluate the performance of installed Photovoltaic (PV) panels against theoretical yield, and retrieve cadastral data for investment appraisal. The tool integrates irradiance models, building geometry data, and shading analysis to produce building-level assessments, supporting both public authorities planning solar roll-out and private investors conducting due diligence.
Two tools that address interoperability at the data modelling level. BrickLLM uses Large Language Models to generate RDF files conforming to the Brick Schema ontology from natural-language descriptions of building systems—dramatically reducing the expert time required to create semantic building data models. The Brick Assistant provides an interactive query interface over Brick-modelled building data. Together, these tools lower the technical barrier to adopting ontology-based data management, a prerequisite for platform interoperability.
A customisable environment for optimising building energy systems using timeseries analytics. The tool enables smart heating control based on weather forecasts, allowing automated temperature setpoint adjustments that minimise energy consumption while maintaining thermal comfort. Integration with building automation systems enables closed-loop control. The tool was validated with data from project partners.
A freely available Python library implementing ISO EN 52016 and other relevant standards for the calculation of building energy needs for heating and cooling, indoor environmental quality, and system behaviour at hourly resolution. Released by Eurac Research under an open licence, pyBuildingEnergy has already attracted community uptake among researchers and is relevant to the implementation of the recast EPBD, which references ISO 52016 as a preferred calculation standard.
Table 2. MODERATE toolkit. (Source: MODERATE project).
MODERATE's tools were not developed in a lab. They were designed with, and validated against, datasets provided by industrial partners managing real building portfolios across Europe:
Across these datasets, synthetic data generation was applied to create openly publishable versions, with demonstrated preservation of statistical properties (correlation structures, marginal distributions, temporal patterns). MODERATE platform hosts a populated dataset catalogue with data handling capabilities, and a selection of synthetic datasets are published on Zenodo with DOIs, enabling reproducible research.
Early adopter feedback collected through structured workshops in 2025 confirmed strong concept validation (mean score 4.4/5.0 for interest and perceived value) and identified tool interoperability as the primary gap to address before full commercial deployment (operational readiness score 3.0/5.0). The gap between interest and readiness reflects not scepticism about the platform's value but the practical challenges of integrating tools into existing organisational IT workflows—a challenge the project is actively addressing in its final phase.
A recurring failure mode of EU-funded research platforms is that results become inaccessible once project funding ends, locked inside consortium IP agreements or dependent on public co-funding for operation. MODERATE has made structural choices to prevent this:
MODERATE's outputs are directly relevant to the implementation of several EU legislative instruments:
The platform is particularly relevant to less-digitised regions seeking to build data infrastructure for renovation programmes. The modular design means that a regional energy agency or municipality can deploy their own infrastructure at low cost, connecting their building stock data to a shared analytics ecosystem without transferring custody of sensitive data outside their jurisdiction.
The MODERATE project demonstrates that overcoming data fragmentation is both a technical and institutional challenge—and one that can be addressed through open, interoperable, and privacy‑preserving digital infrastructure. By combining federated data custody, synthetic data generation, and a suite of open‑source analytics tools, the platform offers a practical pathway for transforming raw building data into actionable intelligence across the full lifecycle of buildings. Its validation on real portfolios, alignment with EU legislation, and commitment to open science position MODERATE as a durable foundation for data‑driven renovation strategies, digital building logbooks, and emerging energy communities. As the sector moves toward large‑scale implementation of the EPBD and the broader decarbonisation agenda, MODERATE provides a replicable blueprint for regions and organisations seeking to build trustworthy, future‑proof data ecosystems.
Check the MODERATE Project website for more information.
