astro-legacy-archive/quiet-released-products

QUIET released measurement products

[image]

This dataset mirrors a selected eight-file fragment of the QUIET release served by LAMBDA: the four raw maps under its “QUIET Galactic Fields 1st+2nd Season Observations” heading and the four first- and second-season angular-power tables. It is not the complete QUIET release. LAMBDA separately releases raw-map covariance matrices, QUIET+WMAP coadds and their covariance matrices, masks, galaxy-map data, beams, and bandpasses.

Each configuration is named with the corresponding source filename stem. The four map configurations preserve their FITS table structure; the four band-power configurations preserve the distinct published pipeline products.

How to use

python -m venv .venv && .venv/bin/pip install datasets huggingface_hub pyarrow

from datasets import load_dataset
dataset = load_dataset("astro-legacy-archive/quiet-released-products", "QUIET_first_season_bandpowers", split="train[:2]", columns=["lmin", "lmax", "EE", "+dEE", "-dEE"])
for row in dataset:
    print(row["lmin"], row["lmax"], row["EE"], row["+dEE"], row["-dEE"])

25.0 75.0 0.33 0.16 -0.11
76.0 125.0 0.82 0.23 -0.2

import json
from huggingface_hub import hf_hub_download
import pyarrow.parquet as pq
path = hf_hub_download("astro-legacy-archive/quiet-released-products", "quiet_qband_G-1_map_v2.parquet", repo_type="dataset")
print(json.loads(pq.read_metadata(path).metadata[b"healpix_map"])["healpix"])

{'nside': 512, 'npix': 3145728, 'ordering': 'NESTED', 'coordsys': 'G', 'coordsys_source': {'keyword': 'COORDSYS', 'hdu': 'table'}, 'pixel_rule': 'element j of row i is HEALPix pixel i * 1024 + j in the ordering named by ORDERING'}

Raw QUIET maps

The map configurations are quiet_qband_G-1_map_v2, quiet_qband_G-2_map_v2, quiet_wband_G-1_map_v2, and quiet_wband_G-2_map_v2. Each source is an implicit full-sky HEALPix binary table with NSIDE=512, 3,145,728 pixels, NESTED ordering, and Galactic coordinates (COORDSYS='G'). Each Parquet row retains one source FITS row; each of its three fields is a length-1,024 list. Element j of row i is therefore HEALPix pixel i * 1024 + j in the source NESTED ordering.

The three source identifiers and their order are unchanged:

The source headers also record POLCCONV='COSMO' and BAD_DATA=-1.6375e+30. In every selected map, all 3,145,728 values of source field TEMPERATURE are the stored representation of that source BAD_DATA sentinel. The field and every value remain unchanged. Q-band G-1 also contains 46 occurrences per Q/U field of the exact header representation. The embedded healpix-map 2.1.0 document records each representation and its exact count in each source field; no validity column, null, or replacement value is added.

The non-sentinel populations are the same in each map's Q and U fields:

Band-power tables

Every row is one published multipole bin, and every column is float64 in the source column order. The ASCII headers do not declare units. The first-season paper labels its bin-bound column “ℓ bin”; the second-season paper labels the two columns “ℓmin” and “ℓmax”. These dimensionless multipole bounds are recorded separately from the remaining columns, whose table notes give CMB thermodynamic uK^2, scaled as C_l l(l+1)/(2 pi). The structured records preserve the source-header silence and the separately pinned paper passages.

EE, BB, and EB are the published band powers. The signed +dXX and -dXX identifiers distinguish the published upper and lower error bars; they are not replaced by a derived symmetric uncertainty. BB 95% U.L. is the published 95% confidence upper limit on the BB band power. In the second-season tables, EE/q, +dEE/q, and -dEE/q give the EE spectrum shape normalized to q=1. Pipeline B instead publishes one dEE, dBB, and dEB error column for each spectrum.

The first-season source header identifies Pipeline A as the official QUIET result and notes foreground contamination in its lowest EE multipole bin. The separate Pipeline B header repeats that qualification and says Pipeline A's values are official because the systematic studies used Pipeline A. The two second-season files remain separate ML and PCL analysis products, as published by LAMBDA; their source headers associate them with Figures 4–6 and Table 3 of arXiv:1207.5034v1.

Provenance and verification

LAMBDA's QUIET product table is the release index. Its map description documents the HEALPix maps and their CMB polarization convention; its second-season results description documents the ML and PCL columns.

The build pins the byte size and SHA-256 digest of every source before parsing. The four FITS files are each 75,507,840 bytes. Their respective digests, in the map configuration order above, are e8af7ca5aa06add89b964116cff0d2fb154a9d38d04f39696e2e564621013a29, cbc668753b8ceb8b08980302ea65102c3e9ecdbea3533b1e0922d0d800a97672, bdd394d252a86b215d3cf6fc332ad61f7d5fec0948620d8fe594854d249f5fb7, and 8fcb6007437512553e0c2378c018d396c8dda715d2a2cc8121633cf364167aa0.

The four ASCII tables are respectively 1,040, 989, 1,693, and 1,694 bytes, with SHA-256 digests 1d539c4b8e74a84ee79bbfbe15b3edd4cf9cad84964dbe235317d4ed251264c9, dfcfa6180e6dd8becddc3742e26242b5426e0a92ce1d62ea0297a12d2c387f11, 0eb688100af3089096d6d63429c2e07c9383814359df8cb38419b85ef6bffcff, and 80520c6ffc8eddf8d99adc4e88a8e44f6aa4a32f51797f283c016b33c9406c38.

Map verification compares the selected FITS binary-table columns with the Parquet arrays in source column and row order, including each field's shape, dtype, and value bits. The current live FITS objects end at the FITS block boundary; the verification promise covers the source table data stream and identifiers, not FITS container serialization. Table verification compares every parsed Arrow array with the Parquet arrays. The original ASCII header bytes are embedded verbatim with their encoding recorded: Pipeline B is ISO-8859-1 and the other three tables are ASCII. Whitespace and decimal formatting outside those headers are not reconstructed from the typed values.

The embedded metadata records source URLs, sizes, digests, complete source headers, field descriptions, and the repository, commit, working directory, and command needed to regenerate the files. Maps carry healpix-map schema 2.1.0; band-power tables carry statistical-product schema 1.5.0.

Licensing and credit

No explicit licence is published on the checked LAMBDA QUIET product pages, so the dataset retains license: unknown. Separately, HEASARC's first-party data policy states: “HEASARC materials are all available freely for your use.” This archive-wide use statement is not treated as an invented named licence. LAMBDA requests the following acknowledgement when research benefits from its use: “We acknowledge the use of the Legacy Archive for Microwave Background Data Analysis (LAMBDA), part of the High Energy Astrophysics Science Archive Center (HEASARC). HEASARC/LAMBDA is a service of the Astrophysics Science Division at the NASA Goddard Space Flight Center.” The request is published on LAMBDA's acknowledgement page. The exact papers named by the source headers are arXiv:1012.3191 for the first-season Pipeline A table and arXiv:1207.5034v1 for both second-season tables; these links are also structured in those artifacts.

Citation

Please cite this dataset:

@misc{https://doi.org/10.57967/hf/10434,
  doi = {10.57967/hf/10434},
  url = {https://huggingface.co/datasets/astro-legacy-archive/quiet-released-products},
  author = {{Astro Legacy Archive}},
  title = {quiet-released-products},
  publisher = {Hugging Face},
  year = {2026}
}

The initial public release is Hub revision `78915f2f97b1fc267619d119bcb08a3ea26c3880`.

Errata and requests

To report an erratum, propose a correction, or ask a question about this dataset, open a Discussion or Pull Request in its Community tab; posting requires a free Hugging Face account. Dataset requests and archive-wide questions may be posted in the Community tab of any dataset. Confirmed corrections are made in the source conversion tools and published by rebuilding the dataset.