Unusual research data from our everyday world has been added to the ARP Data Repository in an unusual way and is available to anyone interested.
Lilla Barancsuk, Dalma Günter, Veronika Oláhné Groma, and Bálint Sinkovics, members of the Energy Strategy and Environmental Effects Research Group (HUN-REN Center for Energy Research, Institute of Technical Physics and Materials Science), have made a series of images of the sky, consisting of hundreds of thousands of files, available in the ARP Data Repository.
An important goal of the research group’s work is to develop a reliable, ultra-short-term solar power generation estimation method capable of forecasting expected global radiation and photovoltaic power generation. Since production is primarily influenced by highly variable cloud cover, the researchers capture and analyze a series of sky images taken at extremely high intervals for their estimates. A wide-angle sky camera and an associated weather station are available for the studies at the Energy Research Center’s facility.
During the research conducted over the past few years, a massive dataset of several hundred gigabytes was generated; the images and accompanying measurements were processed using artificial neural networks. Lilla Barancsuk, ARP Ambassador, made hundreds of thousands of image files created between November 19, 2021, and August 4, 2024, available through the HUN-REN Data Repository Platform with the support of the ARP 2024 Ambassador Program.
The size of the data package containing the images posed significant challenges for the ARP platform’s infrastructure. The solution was to develop an alternative multi-tier storage system in which a dedicated data package was created for the entire research material (by packaging the sky image sets on a daily basis), and separate data packages were created for each month’s sky images, available in a previewable format.
To automate the upload process, the research team used a Python-based script they developed themselves.
Hosting and making the sky images available marks an important milestone for the research team. It did not fit within the capabilities of other previously explored data repositories, but the ARP team made it possible to store and make this vast and valuable dataset accessible through a customized solution.
With the developed solution, the research data is now easy to manage and can be easily discovered, accessed, and downloaded by other researchers. The December sky images rank among ARP’s most popular research data; as of November 2025, the files in the dataset have been downloaded more than 46,500 times.
Data packages are available:
Ground-based all-sky imagery and weather data
All-sky imagery for the month of 2021/11
All-sky imagery for the month of 2021/12
Image ource: Lilla Barancsuk (HUN-REN EK)