{"success":true,"data":{"author":"NASA-IBM AI4Science (Surya / SuryaBench team)","bucketAccessType":"public","bucketName":"nasa-ibm-ai4science/surya-bench-flare-forecasting","bucketPrefix":"","category":"Astronomy & Space","cloudProvider":"huggingface","createdAt":"2026-11-16T02:38:29.750Z","description":"Hourly labels for full-disk solar flare forecasting from NOAA GOES X-ray flare event lists, the flare task of SuryaBench, the benchmark suite for the NASA-IBM Surya heliophysics foundation model. Each timestamp from 13 May 2010 to 31 December 2024 opens a 24-hour rolling prediction window annotated with the maximum GOES class (peak 1-8 Angstrom X-ray flux, or FQ for flare-quiet), a cumulative flare index summing C, M and X subclasses with weights 1, 10 and 100, and two binary targets: label_max (window contains a flare of M1.0 or above) and label_cum (cumulative index of 10 or more). The labels support binary, ordinal (flare class) and regression (cumulative index) formulations and are meant to be paired with full-disk SDO AIA and HMI inputs such as the Surya core stack on s3://nasa-surya-bench. CSV splits follow the shared SuryaBench calendar: train (15 February to 31 December of 2010 to 2019), validation (15 to 31 January), leaky_validation (1 to 14 January and 1 to 14 February) and test (2020 to 2024). Small (a few MB); load with pandas or Hugging Face datasets. CC-BY-4.0 as stated in the card.","displayTitle":"SuryaBench Full-Disk Solar Flare Forecasting Labels: Hourly 24-Hour GOES X-ray Flare Windows (2010-2024)","id":"f978e621-2f10-5b44-977a-64e07f967739","license":"CC-BY-4.0","licenseEvidenceUrl":"https://huggingface.co/datasets/nasa-ibm-ai4science/surya-bench-flare-forecasting","objectCountEstimate":7,"probeStatus":"ok","referenceUrl":"https://github.com/NASA-IMPACT/Surya","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"nasa-ibm-ai4science/surya-bench-flare-forecasting\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"nasa-ibm-ai4science/surya-bench-flare-forecasting\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=nasa-ibm-ai4science, license:cc-by-4.0, size_categories:100K<n<1M, format:csv, downloads=333"],"format":"mixed","generatedAt":"2026-11-16T02:38:35.966Z","layout":{"example":"data.csv","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"timestamp,max_goes_class,cumulative_index,label_max,label_cum\n2010-05-13 00:00:00,B2.8,0.0,0,0\n2010-05-13 01:00:00,B2.8,0.0,0,0\n2010-05-13 02:00:00,B2.8,0.0,0,0\n2010-05-13 03:00:00,B2.8,0.0,0,0\n2010-05-13 04:00:00,FQ,0.0,0,0\n2010-05-13 05:00:00,FQ,0.0,0,0\n2010-05-13 06:00:00,FQ,0.0,0,0\n2010-05-13 07:00:00,FQ,0.0,0,0\n2010-05-13 08:00:00,FQ,0.0,0,0\n2010-05-13 09:00:00,FQ,0.0,0,0\n2010-05-13 10:00:00,FQ,0.0,0,0\n2010-05-13 11:00:00,FQ,0.0,0,0\n2010-05-13 12:00:00,FQ,0.0,0,0\n2010-05-13 13:00:00,FQ,0.0,0,0\n2010-05-","encoding":"utf8","key":"data.csv","truncatedAt":512}],"sampleObjects":["data.csv","leaky_validation.csv","test.csv","train.csv","validation.csv"],"sidecars":[{"bytes":"---\nlicense: cc-by-4.0\ntags:\n- flare\n- space-weather\n- GOES-flare\n- binary-classification\n- forecasting\nconfigs:\n- config_name: default\n  default: true\n  data_files:\n  - split: train\n    path: train.csv\n  - split: validation\n    path: validation.csv\n  - split: test\n    path: test.csv\n  - split: leaky_validation\n    path: leaky_validation.csv\n---\n\n# Full-disk Solar Flare Forecasting Dataset\n\n## Dataset Summary\n\nThis dataset provides labels for solar flare forecasting derived from NOAA GOES flare events from May 2010 to December 2024. Labels are constructed using a 24h rolling prediction window sampled at an hourly cadence. Each window is annotated with both **max GOES class** (based on peak X-ray flux) and **cumulative flare index**.  \n\nTwo derived binary labels are included for forecasting tasks:  \n\n- **`label_max`**: 1 if the maximum flare intensity in the window is ≥ M1.0.  \n- **`label_cum`**: 1 if the cumulative flare intensity in the window is ≥ 10.  \n\nFor completeness, we also include (1) max GOES class, which is determined from the peak X-ray flux of the most intense flare in the prediction window; and (2) cumulative index determined from all ≥C-class flares in the prediction window.\n\n\n## Supported Tasks and Applications\n\n- `binary-classification`: Predict whether a time window will contain significant flaring activity.  \n`ordinal-classification`: Predict flare-class of a given instance.  \n- `regression`: Predict cumulative flare index of a given instance.  \n\n\n## Dataset Structure\n\n### Data Files\n\n- `train.csv`: Instances from Feb 15 to Dec 31 in each year between 2010–2019  \n- `validation.csv`: Instances from Jan 15–31 of each year between 2010–2019  \n- `test.csv`: All instances from each year between 2020–2024  \n- `leaky_validation.csv`: Instances from Jan 1–14 and Feb 1-14 of each year between 2010–2019\n\n### Features\n\nEach record includes four label fields:  \n- **`max_goes_class`**: Maximum GOES flare class (e.g., C5.2, M1.0, X3.2) in the prediction window, or `FQ` if no flares are present.  \n- **`cumulative_index`**: Weighted sum of flare subclasses ≥C-class in the prediction window.  \n  - C-class contributes weight ×1, M-class ×10, X-class ×100.  \n  - For example, an M2.0 flare adds 20, while an X3.5 flare adds 350.  \n- **`label_max`**: Binary label, 1 if `goes_class` ≥ M1.0, else 0.  \n- **`label_cum`**: Binary label, 1 if `cumulative_index` ≥ 10, else 0.  \n\nExample entry (in JSON format):  \n\n```json\n{\n  \"timestamp\": \"2011-02-14 03:00:00\",\n  \"goes_class\": \"X2.2\",\n  \"cumulative_index\": 297.1,\n  \"label_max\": 1,\n  \"label_cum\": 1\n}\n```\n\n## Dataset Details\n\n| Field | Description |\n|------------------------|---------------------------------------------|\n| **Temporal Coverage** | May 13, 2010 – Dec 31, 2024 |\n| **Data Format** | CSV (.csv), string-based schema |\n| **Data Shape** | (1, 4) per instance |\n| **Data Size** | Total 128,328 instances |\n| **Cadence** | 1 hour |\n| **Total File Size** | ~4.4MB |\n\n## License\nThis dataset is licensed under the Creative Commons Attribution 4.0 International (CC BY 4.0) License.\n\n## Authors\n- Jinsu Hong, [jhong36@gsu.edu](mailto:jhong36@gsu.edu)\n- Kang Yang, [kyang30@gsu.edu](mailto:kyang30@gsu.edu)\n- Berkay Aydin, [baydin2@gsu.edu](mailto:baydin2@gsu.edu)","encoding":"utf8","key":"README.md","name":"README.md","size":3284,"truncated":false}],"sourceObjectKey":"data.csv","starterCell":"import pandas as pd\ndf = pd.read_csv(\"hf://datasets/nasa-ibm-ai4science/surya-bench-flare-forecasting/data.csv\")\ndf.head()","starterSamplePath":"data.csv","totalObjects":7},"summary":"Hourly labels for full-disk solar flare forecasting from NOAA GOES X-ray flare event lists, the flare task of SuryaBench, the benchmark suite for the NASA-IBM Surya heliophysics…","title":"SuryaBench Full-Disk Solar Flare Forecasting Labels: Hourly 24-Hour GOES X-ray Flare Windows (2010-2024)","updatedAt":"2026-11-16T16:57:40.340Z","visibility":"public","websiteUrl":"https://huggingface.co/datasets/nasa-ibm-ai4science/surya-bench-flare-forecasting"}}