Switzerland’s energy transition is accelerating, but planners and analysts have long faced fragmented or incomplete data on how rooftop photovoltaics, heat pumps, batteries, and electric vehicles may spread across the country. Our recent publication introduces a new open dataset that addresses this gap, where future distributed energy resources (DERs) are expected to be deployed at over two million connection points. The dataset can be used to assess grid reinforcement needs, quantify flexibility, support policy and market design, and provide a library of realistic case studies for benchmarking models or developing data-driven applications.
What Is this Dataset and Why It Matters for the Energy Transition
The Swiss energy system is undergoing rapid changes and distributed energy resources (DERs) like rooftop PV, electric vehicles (EVs), batteries (BESS), and heat pumps (HPs) are no longer niche. To plan ahead and ensure a reliable electricity supply, it is crucial not only to know how many of these devices may be installed, but also where they will be located and how they will interact with the grid.
Until now, the available data has been fragmented: some detailed information exists for individual technologies, but it is not harmonised and often inconsistent with national decarbonisation targets. As a result, modelers have been forced to create their own assumptions and scenarios, which limit comparability and coherence.
With our new dataset, these pieces are harmonised into a consistent, high-resolution framework, where DERs are integrated into distribution grids, offering an integrated view of future distributed resources across Switzerland for the first time. Based on official national projections, we provide projected deployment data in Switzerland for the years of 2030, 2040, and 2050, that describe how DERs may be deployed and, importantly, the flexibility they may provide.
Building the Dataset: A Puzzle with Missing Pieces
The challenge in creating a nationwide, high-resolution dataset of DERs lies in combining different, often incomplete sources. For example, data for rooftop PV potential, building archetypes, mobility data, and demand profiles all came in different formats, scales, and levels of detail.
To fill the gaps, we synthesised newly created and existing datasets from various official national sources (see Figure 1 for an overview). Rooftop PV data came from high-resolution maps of solar potential for each Swiss building; battery storage systems were co-located with a share of PV systems following national projections. For heat pumps, we combined the Swiss Building and Dwelling Register with building archetypes describing insulation and heating needs, generating hourly profiles using local weather data. EV charging and flexibility patterns were derived from nationwide mobility simulations covering over three million vehicles. These heterogeneous data streams were processed into a single framework linked to distribution grid models, providing realistic hourly profiles and flexibility attributes for all distribution nodes in Switzerland.
All these data were then harmonised into a single, consistent framework and linked to Swiss distribution grid models. The result is a modular dataset of over two million connection points, with hourly time series and flexibility parameters for each DER type across three years (2030, 2040, and 2050).
Three Snapshots: 2030, 2040, 2050
The dataset includes projections for three years — 2030, 2040, and 2050 — to reflect how Switzerland is expected to transition to full-scale electrification according to the official Swiss energy policy (Energy Perspectives 2050+). Each year represents a distinct phase of this transition:
- 2030 – the first wave: PV on a quarter of roofs, 15% of cars electric, one in three homes with a heat pump.
- 2040 – the acceleration: more than half of roofs and cars, better insulation, and batteries becoming common.
- 2050 – the full scale: almost every roof with PV, all cars electric, most homes heated with a heat pump, and widespread storage (See Figure 2 for a data visualisation)
The intermediate snapshots are just as important as the 2050 target: they show how quickly technologies scale up, allowing us to determine when bottlenecks may appear, and which measures need to be taken along the way, before the 2050 goals are reached.
What We Can (and Cannot) Do with This Data
The dataset can be utilised by researchers and professionals to assess grid reinforcement needs, quantify the flexibility potential of DERs, or test how much storage could help mitigate seasonal mismatches. It is also valuable for policy and market design, such as evaluating incentive schemes for demand response or comparing tariff and regulatory structures. Moreover, it can be used as a library of realistic case studies to benchmark models, develop new optimization methods, or train machine learning tools without exposing sensitive real-world data.
A key feature of the dataset is that it not only provides generation and demand profiles but also detailed flexibility parameters for each controllable technology. PV system data include power limits; batteries are characterised by their rated power, energy capacity, and round-trip efficiency; heat pumps include the thermal properties, and comfort bounds of the buildings they serve; and EVs feature upper and lower charging profiles with daily shiftable energy. These parameters allow users to decide on the operational strategy according to their objectives.
Since real-world information on distribution grids and households is sensitive and fragmented, our dataset is synthesised from public sources and projections, which means it should be viewed as realistic data rather than real data. Because of this, the dataset is not suitable for analysing very local issues, such as checking whether a specific feeder in one neighborhood will be overloaded. Instead, its strength lies in large-scale studies where consistency and representativeness matter more than the exact location of individual devices.
Accessing the Data
The dataset is openly available on Zenodo, along with documentation. The methodology and the structure of the dataset are detailed in the relative publication in Nature Scientific Data. The codebase is public on GitHub.
