{"success":true,"data":{"author":"nyu-planet-finders, from TESS-SPOC (NASA Ames) via MAST","bucketAccessType":"public","bucketName":"nyu-planet-finders/tess-spoc-ffi-binned","bucketPrefix":"","category":"Astronomy & Space","cloudProvider":"huggingface","createdAt":"2026-11-21T20:46:01.522Z","description":"Every TESS-SPOC full-frame-image (FFI) light curve of the Transiting Exoplanet Survey Satellite (TESS) for Sectors 1 to 81, reduced to 30-minute bins for machine-learning searches for transiting exoplanets. The source is the TESS Science Processing Operations Center (SPOC) high-level science product at MAST (Caldwell et al. 2020, DOI 10.17909/t9-wpz1-8s54). Processing removed cadences flagged by lightkurve's default quality bitmask, normalized PDCSAP flux, flux error and SAP background by median flux, subtracted median centroids, and averaged into 30-minute bins with empty bins omitted. One Parquet file per sector (s0001 to s0081, 2 to 4 GB each, 252.8 GB in total) holds one row per star per sector, 12,797,720 rows (about 160,000 stars per sector), with TIC ID, sector, TESS magnitude, effective temperature and radius, and aligned list columns of time (BTJD), flux, flux error, centroid offsets, background and points per bin. Read row groups with pyarrow or the Hugging Face file system. Built by nyu-planet-finders for their transit detector. TESS-SPOC HLSP products are CC BY 4.0 per MAST; acknowledge TESS, SPOC and MAST.","displayTitle":"TESS-SPOC Full-Frame-Image Light Curves Binned to 30 Minutes: 12.8 Million Star-Sectors, Sectors 1-81 (Transit Search)","huggingFaceUrl":"https://huggingface.co/datasets/nyu-planet-finders/tess-spoc-ffi-binned","id":"5c3d7f7e-ac78-5615-b50c-204347de9dbb","license":"CC-BY-4.0","licenseEvidenceUrl":"https://archive.stsci.edu/hlsp/tess-spoc","objectCountEstimate":83,"probeStatus":"ok","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"nyu-planet-finders/tess-spoc-ffi-binned\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"nyu-planet-finders/tess-spoc-ffi-binned\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=nyu-planet-finders, license:cc-by-4.0, size_categories:10M<n<100M, format:parquet, downloads=1033"],"format":"mixed","generatedAt":"2026-11-21T20:48:55.288Z","layout":{"example":"s0001.parquet","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"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","encoding":"hex","key":"s0001.parquet","truncatedAt":512}],"sampleObjects":["s0001.parquet","s0002.parquet","s0003.parquet","s0004.parquet","s0005.parquet"],"sidecars":[{"bytes":"---\nlicense: cc-by-4.0\npretty_name: TESS-SPOC full-frame-image light curves, binned to 30 minutes\ntags:\n- astronomy\n- exoplanets\n- tess\n- light-curves\n- time-series\nsize_categories:\n- 10M<n<100M\nconfigs:\n- config_name: default\n  data_files:\n  - split: train\n    path: \"s*.parquet\"\n---\n\n# TESS-SPOC full-frame-image light curves, binned to 30 minutes\n\nEvery TESS-SPOC full-frame-image (FFI) light curve of Sectors 1-81, reduced to 30-minute bins for machine-learning\ntransit searches. This is the data the nyu-planet-finders detector searches.\n\n- One Parquet file per sector (`s0001.parquet` ... `s0081.parquet`)\n- One row per star per sector: 12,797,720 rows, about 160,000 stars per sector, 252.8 GB\n- `split: train` only means \"all rows\"; this repository is not divided into training and test sets\n\n## Columns\n\n| Column | Type | Description |\n|---|---|---|\n| `tic_id` | int64 | TESS Input Catalog ID |\n| `sector` | int16 | TESS sector |\n| `tmag` | float64 | TESS magnitude (from the light-curve header) |\n| `teff` | float64 | Stellar effective temperature [K]; may be missing |\n| `radius` | float64 | Stellar radius [R_sun]; may be missing |\n| `time` | list[float64] | Bin centers, BTJD (BJD - 2457000) |\n| `flux` | list[float32] | PDCSAP flux divided by its median, averaged into 30-minute bins |\n| `flux_err` | list[float32] | Uncertainty of each binned flux value, same units |\n| `centr_x`, `centr_y` | list[float32] | Flux-weighted centroid minus its median [pixels, about 21 arcsec]; NaN in a bin with no valid centroid |\n| `bkg` | list[float32] | SAP background divided by the star's median flux, binned like the flux |\n| `n_points` | list[int16] | Number of original cadences averaged into each bin |\n\nThe list columns of a row are aligned element by element; their length is the number of non-empty bins (about\n950-1,300 per sector).\n\n## Processing\n\n1. Read every `*_tess_v1_lc.fits` file of the TESS-SPOC high-level science product from the MAST public S3\n   bucket (`s3://stpubdata/mast/hlsp/tess-spoc/`).\n2. Removed cadences flagged by lightkurve's default TESS quality bitmask (17087), and cadences without a\n   finite time or PDCSAP flux. A cadence with good flux but no centroid is kept.\n3. Divided flux, flux error and background by the median PDCSAP flux; subtracted the median centroid.\n4. Averaged into 30-minute bins. Empty bins are omitted, so gaps appear as missing time.\n\nThe FFI cadence changed during the mission, so `n_points` is about 1 for Sectors 1-26 (30-minute images),\nabout 3 for Sectors 27-55 (10-minute) and about 9 for Sectors 56-81 (200-second).\n\nDifferences from [tess-spoc-2min-binned](https://huggingface.co/datasets/nyu-planet-finders/tess-spoc-2min-binned):\nthe 2-minute data dropped every flagged cadence (QUALITY == 0) and every cadence without a centroid, and has no\n`bkg` or `n_points` columns. The original quality flags are not stored.\n\n## Usage\n\n```python\nimport pyarrow.parquet as pq\nfrom huggingface_hub import HfFileSystem\n\nfs = HfFileSystem()\nwith fs.open(\"datasets/nyu-planet-finders/tess-spoc-ffi-binned/s0040.parquet\", \"rb\") as handle:\n    table = pq.ParquetFile(handle).read_row_group(0)    # about 2,000 stars; read one row group at a time\nrow = table.slice(0, 1).to_pylist()[0]\n```\n\nEach file is 2-4 GB. Read row groups rather than whole files on machines with little memory. The Hugging Face\nviewer shows only the first few hundred elements of each list.\n\n## Derived datasets\n\n[lc-pretrain](https://huggingface.co/datasets/nyu-planet-finders/lc-pretrain) puts these light curves, with the\n2-minute and Kepler data, on one global 30-minute grid with noise normalization, quality flags and splits by\nstar; [lc-finetune](https://huggingface.co/datasets/nyu-planet-finders/lc-finetune) adds transit labels.\n\n## Source and acknowledgment\n\nTESS-SPOC FFI light curves (Caldwell et al. 2020, RNAAS 4, 201; doi:10.17909/t9-wpz1-8s54), produced by the\nTESS Science Processing Operations Center at NASA Ames and obtained from the Mikulski Archive for Space\nTelescopes (MAST) at the Space Telescope Science Institute. The TESS-SPOC products are re","encoding":"utf8","key":"README.md","name":"README.md","size":4258,"truncated":true}],"sourceObjectKey":"s0001.parquet","starterCell":"from datasets import load_dataset\nds = load_dataset(\"nyu-planet-finders/tess-spoc-ffi-binned\")\nprint(ds)","starterSamplePath":"s0001.parquet","totalObjects":83},"summary":"Every TESS-SPOC full-frame-image (FFI) light curve of the Transiting Exoplanet Survey Satellite (TESS) for Sectors 1 to 81, reduced to 30-minute bins for machine-learning searches…","title":"TESS-SPOC Full-Frame-Image Light Curves Binned to 30 Minutes: 12.8 Million Star-Sectors, Sectors 1-81 (Transit Search)","updatedAt":"2026-11-22T01:02:44.364Z","visibility":"public","websiteUrl":"https://huggingface.co/datasets/nyu-planet-finders/tess-spoc-ffi-binned"}}