Photo-z Server Data¶
Back to Photo-z Server documentation
Datasets classification¶
Official Datasets
The official datasets will be produced by Rubin's Data Management team and will be available to the LSST Community through the Photo-z Server. These datasets will include reference redshift catalogs, training sets, and photo-z estimates associated with the LSST data releases. The datasets will be released in a phased manner, starting with the first data release (DR1) and continuing with subsequent releases as the LSST survey progresses.
For now, the Rubin Observatory PZ Data Products page is empty.
Unofficial Datasets
Besides data uploaded by users, the User-generated Data Products page hosts set of datasets that were prepared by the LIneA team for educational purposes, to serve as use case examples for the Photo-z Server tutorials. In addition, the DP1 datasets generated by PZ Science Unit from Rubin’s Commissioning Team, described in the tech note SITCOMTN-154 are also available there. All these datasets are not classified as Official Datasets.
Data Preview 1¶
ATTENTION: Preliminary Datasets
These datasets were produced by the PZ Science Unit — a working group from Rubin’s Commissioning Team — during the Initial studies of photometric redshifts with LSSTComCam from DP1. All results, along with detailed dataset descriptions, are available in the tech note SITCOMTN-154.
These datasets are NOT classified as Official Datasets released by Rubin's DM team.
The datasets described in the tech note are available in the Photo-z Server as data products with SITCOMTN-154
suffix. Their links and short description are organized in product types below.
Object Catalogs¶
Data products containing object tables described in Section 2.1 and listed in Table 1 from the tech note SITCOMTN-154.
Data Product | Data set | Selection | Number of objects |
---|---|---|---|
DP1 (available in the RSP) | Complete DP1 Object Catalog | None | 2,299,757 |
ECDFS+EDFS+SV_95 gold SITCOMTN-154 | ECDFS+EDFS+SV_95 | gold | 375,610 |
SV_38 gold_4_band SITCOMTN-154 | SV_38 | gold_4_band | 169,034 |
All DP1 fields comprehensive Gold dataset beyond the fields ECDFS+EDFS+SV_95 and SV_38, where spectroscopic data are available (dataset not listed in Table 1 from the tech note SITCOMTN-154):
Data Product | Data set | Selection | Number of objects |
---|---|---|---|
DP1 Gold all SITCOMTN-154 | All fields | gold | 686,334 |
Reference Redshift Catalogs¶
Reference Redshift Catalogs from Individual Surveys¶
Data products containing reference redshifts catalogs (before matching with DP1 Object table) separated by the origin survey as listed in Table 2 from the tech note SITCOMTN-154.
Note: These datasets are already filtered to the ECDFS field and cleaned with the selection criteria described in Section 2.2.1 from the tech note SITCOMTN-154. For the complete original catalogs, please go to the External datasets section below.
Additional spectroscopic data from DESI DR1 were used as an independent test set for validating the photo-z estimates. Since DESI DR1 is a very large dataset that extends beyond the DP1 footprint, it was filtered to include only the ECDFS field.
Data Product | Reference | Number of Redshifts |
---|---|---|
DESI DR1 inside DP1 footprint | DESI Collaboration et al. (2025) | 50,634 |
Combined Redshift Catalog¶
A single file containing all reference redshifts combined from the individual surveys listed above (excluding DESI), as described in Section 2.2.1 of the technical note SITCOMTN-154.
Data Product | Number of Redshifts |
---|---|
ComCam ECDFS z catalog SITCOMTN-154 | 104,070 |
Training and Test Sets¶
On the Photo-z Server, the product type "Training Set" comprehends all samples resulting from the matching between a reference redshift and an object catalog. That might include training and test sets together in a same file or independent sub-samples uploaded separately. For the latter, both training and test sets are tagged as "Training Set".
Data products containing training and test sets generated from the ComCam ECDFS z catalog listed in Table 1 from the tech note SITCOMTN-154 are:
Data Product | Data set | Selection | Number of objects |
---|---|---|---|
training_v1 match_prelim SITCOMTN-154 | training_v1 | match_prelim | 7,000 |
test_v1 match_prelim SITCOMTN-154 | test_v1 | match_prelim | 2,437 |
training_v4 match_ecdfs SITCOMTN-154 | training_v4 | match_ecdfs | 6,778 |
test_v4 match_ecdfs SITCOMTN-154 | test_v4 | match_ecdfs | 2,905 |
test_DESI match_desi SITCOMTN-154 | test_DESI | match_desi | 2,728 |
Training Results¶
Estimator data models listed in Table 7 and described in the Appendix A.3 from the tech note SITCOMTN-154.
Configuration Files
As mentioned in Section 3.4 and Appendix A.1 from the tech note SITCOMTN-154, the configuration files dp1.yaml (complete set of configurations tested, labeled as analysis flavors) and dp1_v4.yaml (optimized configuration parameters) are available on the GitHub repository rail_project_config
.
Validation Results¶
Photo-z point estimates, QP Ensaembles, and evaluation metrics related to the results shown in Table 4. Files uploaded from directories listed in Table 7 and described in the Appendix A.4 from the tech note SITCOMTN-154.
Photo-z Estimates¶
Photo-z Tables¶
PZ tables produced as part of the initial studies with commissioning data described in SITCOMTN-154. Data uploaded from the directories listed in Table 7. For more datasets from the tech note, visit https://docs.linea.org.br/en/data/pz_server_data.html
Data Product | Number of objects |
---|---|
PZ table dp1_all gold baseline SITCOMTN-154 | 686,334 |
PZ table dp1_all gold dp1_optimize | ⏳ |
PZ table dp1_all gold dp1_optimize 4band | ⏳ |
PZ table dp1_sv38 gold baseline SITCOMTN-154 | 169,034 |
PZ table dp1_sv_38 gold dp1_optimize | ⏳ |
PZ table dp1_sv_38 gold dp1_optimize 4band | ⏳ |
Data Product | Number of objects |
---|---|
PZ table DESI gold baseline | 2728 |
QP Ensembles¶
QP Ensables for the photometric sets of DP1 are not lightweight files, so they are not uploaded to the Photo-z Server. However, they are distributed via LSDB, USDF, and NERSC, as described in Appendix B.1, B.3, and B.4 respectively.
LSDB.io at LIneA
Large datasets including photo-z results will be made available at LSDB.io at LIneA service soon.
External datasets¶
Public data collected from the literature and hosted on the Photo-z Server.
ATTENTION: External Datasets
⏳ Documentation in preparation.
Back to Photo-z Server documentation