The TIGER/Line shapefiles and related database files (.dbf) are an extract of selected geographic and cartographic information from the U.S. Census Bureau's Master Address File / Topologically Integrated Geographic Encoding and Referencing (MAF/TIGER) Database (MTDB). The MTDB represents a seamless national file with no overlaps or gaps between parts, however, each TIGER/Line shapefile is designed to stand alone as an independent data set, or they can be combined to cover the entire nation. Edge refers to the linear topological primitives that make up MTDB. The All Lines Shapefile contains linear features such as roads, railroads, and hydrography. Additional attribute data associated with the linear features found in the All Lines Shapefile are available in relationship (.dbf) files that users must download separately. The All Lines Shapefile contains the geometry and attributes of each topological primitive edge. Each edge has a unique TIGER/Line identifier (TLID) value.

Abstract

👂💉 EHRSHOT is a dataset for benchmarking the few-shot performance of foundation models for clinical prediction tasks. EHRSHOT contains de-identified structured data (e.g., diagnosis and procedure codes, medications, lab values) from the electronic health records (EHRs) of 6,739 Stanford Medicine patients and includes 15 prediction tasks. Unlike MIMIC-III/IV and other popular EHR datasets, EHRSHOT is longitudinal and includes data beyond ICU and emergency department patients.

⚡️Quickstart 1. To recreate the original EHRSHOT paper, download the EHRSHOT_ASSETS.zip file from the "Files" tab 2. To work with OMOP CDM formatted data, download all the tables in the "Tables" tab

⚙️ Please see the "Methodology" section below for details on the dataset and downloadable files.

Methodology

1. 📖 Overview

EHRSHOT is a benchmark for evaluating models on few-shot learning for patient classification tasks. The dataset contains:

• **6,739 **patients
• 41.6 million clinical events
• 921,499 visits
• 15 prediction tasks

%3C!-- --%3E

2. 💽 Dataset

EHRSHOT is sourced from Stanford’s STARR-OMOP database.

• Data follows the OMOP CDM and is fully de-identified.
• Unlike most other EHR research datasets, EHRSHOT is not restricted to ED/ICU visits and instead includes longitudinal patient data for all hospital encounter types.
• EHRSHOT does not contain clinical notes or images.

%3C!-- --%3E

We provide two versions of the dataset:

• EHRSHOT-Original is the same exact dataset used in the original EHRSHOT paper.
• EHRSHOT-OMOP is a more complete version of the EHRSHOT dataset which includes all OMOP CDM tables and additional OMOP metadata.

%3C!-- --%3E

To access the raw data, please see the "Tables" and "Files"** **tabs above:

3. 💽 Data Files and Formats

We provide EHRSHOT in two file formats:

• OMOP CDM v5.4
• Medical Event Data Standard (MEDS)

%3C!-- --%3E

Within the "Tables" tab...

1. %3Cu%3EEHRSHOT-OMOP%3C/u%3E

* Dataset Version: EHRSHOT-OMOP

* Notes: Contains all OMOP CDM tables for the EHRSHOT patients. Note that this dataset is slightly different than the original EHRSHOT dataset, as these tables contain the full OMOP schema rather than a filtered subset.

Within the "Files" tab...

1. %3Cu%3EEHRSHOT_ASSETS.zip%3C/u%3E

* Dataset Version: EHRSHOT-Original

* Data Format: FEMR 0.1.16

* Notes: The original EHRSHOT dataset as detailed in the paper. Also includes model weights.

2. %3Cu%3EEHRSHOT_MEDS.zip%3C/u%3E

* Dataset Version: EHRSHOT-Original

* Data Format: MEDS 0.3.3

* Notes: The original EHRSHOT dataset as detailed in the paper. It does not include any models.

3. %3Cu%3EEHRSHOT_OMOP_MEDS.zip%3C/u%3E

* Dataset Version: EHRSHOT-OMOP

* Data Format: MEDS 0.3.3 + MEDS-ETL 0.3.8

* Notes: Converts the dataset from EHRSHOT-OMOP into MEDS format via the `meds_etl_omop`command from MEDS-ETL.

4. %3Cu%3EEHRSHOT_OMOP_MEDS_Reader.zip%3C/u%3E

* Dataset Version: EHRSHOT-OMOP

* Data Format: MEDS Reader 0.1.9 + MEDS 0.3.3 + MEDS-ETL 0.3.8

* Notes: Same data as EHRSHOT_OMOP_MEDS.zip, but converted into a MEDS-Reader database for faster reads.

4. 🤖 Model

We also release the full weights of **CLMBR-T-base, **a 141M parameter clinical foundation model pretrained on the structured EHR data of 2.57M patients. Please download from https://huggingface.co/StanfordShahLab/clmbr-t-base

**5. 🧑‍💻 Code **

Please see our Github repo to obtain code for loading the dataset and running a set of pretrained baseline models: https://github.com/som-shahlab/ehrshot-benchmark/

Usage

**NOTE: You must authenticate to Redivis using your formal affiliation's email address. If you use gmail or other personal email addresses, you will not be granted access. **

Access to the EHRSHOT dataset requires the following:

• Verified Affiliation with an **Academic, Government, **o

This is a child item of the USGS Data Release: https://doi.org/10.5066/F7513WCB. These data are part of the Gulf Watch Alaska (GWA) long-term monitoring program, nearshore monitoring component. The dataset has three comma separated values (.csv) file exported from a Microsoft Access relational database. The data consist of: 1) size and disease state of sea stars, 2) counts of sea stars, and 3) taxonomic classification.

Abstract

MedAlign is a benchmark dataset of 983 clinician-curated natural language instructions for EHR data, grounded by 275 longitudinal EHRs. It includes reference responses for 303 instructions and supports evaluation of LLMs on healthcare-specific tasks.

Methodology

**IMPORTANT USAGE NOTE: **MedAlign only includes test set examples. No training examples are provided for fine-tuning models.

1. Overview

MedAlign is a longitudinal EHR benchmark for instruction-following with LLMs. The dataset includes:

• 275 patients
• 46,252 clinical notes
• 128 clinical note types
• 3.6 million clinical events

%3C!-- --%3E

2. EHR Data

EHR data is sourced from Stanford’s STARR-OMOP database. Data are standardized in the OMOP CDM schema and are scrubbed on identifying PHI information. Complete technical details are included in the paper, but key highlights:

• Dates are jittered within patient to conceal real dates (but preserve deltas between dates)
• Data for patients %3E= 90 years old are removed

%3C!-- --%3E

• Unstructured text fields not mappable to OMOP standard concepts are redacted

%3C!-- --%3E

• All clinical note text has been scrubbed of PHI variables using hiding-in-plain-sight (HIPS) Carrell et al. 2013.
• HIV test results are redacted.
• Provider names and NPIs are redacted

%3C!-- --%3E

3. Instruction Following Benchmark

See "medalign_instructions_responses_v1_2.zip" for instructions, responses, and EHR text timelines.

Please see our Github repo to obtain code for loading the dataset.

Usage

Access to the MedAlign dataset requires the following:

• Verified Affiliation (Academic, Government, Industry Research Lab). Please use your verified email address when applying, **do not use gmail or personal emails. **Applications using personal, unverified email addresses will be rejected.
• Encryption Verification / Attestation for Data Storage
• Signing the terms of the MedAlign Data Set License 1.0
• Providing a short description of your intended research use of MedAlign
• CITI Training

%3C!-- --%3E

**These data must remain on your encrypted machine. Redistribution of data is FORBIDDEN and will result in immediate termination of access privileges. **

IMPORTANT NOTES:

• Our policy on derived works aligns with PhysioNet's guidelines, requiring that these artifacts be hosted on Redivis. If you create derived research artifacts based on MedAlign (such as additional annotations or synthetic data), please contact us to discuss hosting arrangements.
• Sending MedAlign data over a non-HIPAA-compliant API is a violation of the DUA.

%3C!-- --%3E

Please allow 7-10 business days to process applications.

The TIGER/Line shapefiles and related database files (.dbf) are an extract of selected geographic and cartographic information from the U.S. Census Bureau's Master Address File / Topologically Integrated Geographic Encoding and Referencing (MAF/TIGER) Database (MTDB). The MTDB represents a seamless national file with no overlaps or gaps between parts, however, each TIGER/Line shapefile is designed to stand alone as an independent data set, or they can be combined to cover the entire nation. Edge refers to the linear topological primitives that make up MTDB. The All Lines Shapefile contains linear features such as roads, railroads, and hydrography. Additional attribute data associated with the linear features found in the All Lines Shapefile are available in relationship (.dbf) files that users must download separately. The All Lines Shapefile contains the geometry and attributes of each topological primitive edge. Each edge has a unique TIGER/Line identifier (TLID) value.

The TIGER/Line shapefiles and related database files (.dbf) are an extract of selected geographic and cartographic information from the U.S. Census Bureau's Master Address File / Topologically Integrated Geographic Encoding and Referencing (MAF/TIGER) Database (MTDB). The MTDB represents a seamless national filewith no overlaps or gaps between parts, however, each TIGER/Line shapefile is designed to stand alone as an independentdata set, or they can be combined to cover the entire nation. The Address Range / Feature Name Relationship File (ADDRFN.dbf) contains a record for each address range / linear feature name relationship. The purpose of this relationship file is to identify all street names associated with each address range. An edge can have several feature names; an address range located on an edge can be associated with one or any combination of the available feature names (an address range can be linked to multiple feature names). The address range is identified by the address range identifier (ARID) attribute that can be used to link to the Address Ranges Relationship File (ADDR.dbf). The linear feature name is identified by the linear feature identifier (LINEARID) attribute that can be used to link to the Feature Names Relationship File (FEATNAMES.dbf).

This database contains the macro-invertebrate data that were collected during the AQEM and STAR projects. Samples were taken in 14 European countries using the multi-habitat-sampling (MHS) method as well as the RIVPACS methodology for selected sites. Taxa were identified to the most precise achievable level. Additionally the database contains information on hydromorphology and environmental parameters. The latter include stressor gradients along which the samples were taken. Supplementary fish, macrophyte and diatom data from the STAR project are separately available and can be linked to the invertebrate database. The AQEM and STAR projects were funded by the EU 5th Framework Programme (FP5). More information on this dataset can be found in the Freshwater Metadatabase - BF11 (http://www.freshwatermetadata.eu/metadb/bf_mdb_view.php?entryID=BF11).

This NCEI accession consists of the Surface Ocean CO2 Atlas Version 2021 (SOCATv2021) data product files. The Surface Ocean CO2 Atlas (SOCAT) documents the increase in surface ocean CO2 (carbon dioxide), a critical measure as the oceans are taking up one quarter of the global CO2 emissions from human activity. SOCAT version 2021 has 30.6 million quality-controlled surface ocean fCO2 (fugacity of CO2) observations with an estimated accuracy of better than 5 μatm and a WOCE flag of 2 (good) from 1957 to 2020 for the global oceans and coastal seas. In addition, 2.1 million values with an estimated accuracy of 5 to 10 μatm are available. During quality control, marine scientists assign a flag to each data set, as well as WOCE flags of 2 (good), 3 (questionable) or 4 (bad) to individual fCO2 values. Data sets are assigned flags of A and B for an estimated accuracy of better than 2 μatm, flags of C and D for an accuracy of better than 5 μatm and a flag of E for an accuracy of better than 10 μatm. Bakker et al. (2016) describe the quality control criteria used in SOCAT versions 3 to 2021. Quality control comments for individual data sets can be accessed via the SOCAT Data Set Viewer (www.socat.info). All data sets, where data quality has been deemed acceptable, have been made public. The main SOCAT synthesis files and the gridded products contain all data sets with an estimated accuracy of better than 5 µatm (data set flags of A to D) and fCO2 values with a WOCE flag of 2. Access to data sets with an estimated accuracy of 5 to 10 (flag of E) and fCO2 values with flags of 3 and 4 is via additional data products and the Data Set Viewer (Table 8 in Bakker et al., 2016). SOCAT publishes a global gridded product with a 1° longitude by 1° latitude resolution. A second product with a higher resolution of 0.25° longitude by 0.25° latitude is available for the coastal seas. The gridded products contain all data sets with an estimated accuracy of better than 5 µatm (data set flags of A to D) and fCO2 values with a WOCE flag of 2. Gridded products are available monthly, per year and per decade. Two powerful, interactive, online viewers, the Data Set Viewer and the Gridded Data Viewer (www.socat.info), enable investigation of the SOCAT synthesis and gridded data products. SOCAT data products can be downloaded. Matlab code is available for reading these files. Ocean Data View also provides access to the SOCAT data products (www.socat.info). SOCAT data products are discoverable, accessible and citable. The SOCAT Data Use Statement asks users to generously acknowledge the contribution of SOCAT scientists by invitation to co-authorship, especially for data providers in regional studies, and/or reference to relevant scientific articles. The SOCAT website (www.socat.info) provides a single access point for online viewers, downloadable data sets, the Data Use Statement, a list of contributors and an overview of scientific publications on and using SOCAT. Automation of data upload and initial data checks allows annual releases of SOCAT from version 4 onwards. SOCAT-based data products are used for quantification of the ocean carbon sink, to estimate ocean acidification, for evaluation of biogeochemical sensor data and to evaluate climate models (CMIP). Since 2013 SOCAT products inform the annual Global Carbon Budget. The annual SOCAT releases are made by the SOCAT scientific community as a Voluntary Commitment for United Nations Sustainable Development Goal 14.3 (Reduce Ocean Acidification) (#OceanAction20464). More broadly the SOCAT releases contribute to UN SDG 13 (Climate Action) and SDG 14 (Life Below Water), and to the UN Decade of Ocean Science for Sustainable Development. Hundreds of peer-reviewed scientific publications and high-impact reports cite SOCAT. The SOCAT community-led synthesis product is a key step in the value chain based on in situ inorganic carbon measurements of the oceans, which provides policy makers with essential information on ocean CO2 uptake in climate negotiations. The global need for accurate knowledge of ocean CO2 uptake and its variation (including ocean acidification) makes sustained funding for in situ surface ocean CO2 observations imperative.

For a number of years there has been a great demand for a high-density catalog of accurate stellar positions and proper motions that maintains a consistent system of reference over the entire sky. The Smithsonian Astrophysical Observatory Star Catalog (SAO; SAO Staff 1966) has partially met those requirements, but its positions brought to current epochs now contain errors on the order of 1 second of arc, plus the proper motions in the SAO differ systematically with one another depending on their source catalogs. With the completion of the Second Cape Photographic Catalogue (CPC2; de Vegt et al. 1989), a photographic survey comparable in density to the AGK3 (Dieckvoss 1975) was finally available for the southern hemisphere. These two catalogs were used as a base and matched against the AGK2 (Schorr & Kohlschuetter 1951-58), Yale photographic zones (Yale Trans., Vols. 11-32), First Cape Photographic Catalogue (CPC1; Jackson & Stoy 1954, 55, 58; Stoy 1966), Sydney Southern Star Catalogue (King & Lomb 1983), Sydney Zone Catalogue -48 to -54 degrees (Eichhorn et al. 1983), 124 meridian circle catalogs, and catalogs of recent epochs, such as the Carlsberg Meridian Catalogue, La Palma (CAMC), USNO Zodiacal Zone Catalog (Douglass & Harrington 1990), and the Perth 83 Catalogue (Harwood [1990]) to obtain as many input positions as possible. All positions were then reduced to the system of the FK4 (Fricke & Kopff 1963) using a combination of the FK4, the FK4 Supplement as improved by H. Schwan of the Astronomisches Rechen-Institut in Heidelberg, and the International Reference Stars (IRS; Corbin 1991), then combined with the CPC2 and AGK3. The total number of input positions from which the ACRS was formed is 1,643,783. The original catalog is divided into two parts. Part 1 contains the stars having better observational histories and, therefore, more reliable positions and proper motions. This part constitutes 78 percent of the catalog; the mean errors of the proper motions are +/-0.47 arcsec per century and +/-0.46 arcsec per century in right ascension and declination, respectively. The stars in Part 2 have poor observational histories and consist mostly of objects for which only two catalog positions in one or both coordinates were available for computing the proper motions. Where accuracy is the primary consideration, only the stars in Part 1 should be used, while if the highest possible density is desired, the two parts should be combined. The ACRS was compiled at the U. S. Naval Observatory with the intention that it be used for new reductions of the Astrographic Catalogue (AC) plates. These plates are small in area (2 x 2 deg) and the IRS is not dense enough. Whereas the ACRS was compiled using the same techniques developed to produce the IRS, it became clear as the work progressed that the ACRS would have applications far beyond its original purpose. With accurate positions and proper motions rigorously reduced to both the FK4 and FK5 (Fricke et al. 1988) systems, it does more than simply replace the SAO. Rather, it provides the uniform system of reference stars that has been needed for many years by those who require densities greater than the IRS and with high accuracy over a wide range of epochs. It is intended that, as additional observations become available, stars will be migrated from Part 2 to Part 1, with the hope that eventually the ACRS will be complete in one part. Additional details concerning the compilation and properties of the ACRS can be found in Corbin & Urban (1989) except that the star counts and errors given here supersede the ones given in 1989. The HEASARC revised this database table in August, 2005, in order to add Galactic coordinates. This is a service provided by NASA HEASARC .

The Washington Double Star Catalog (WDS), maintained by the United States Naval Observatory (USNO), is the world's principal database of astrometric double and multiple star information. The WDS Catalog contains positions, discoverer designations, epochs, position angles, separations, magnitudes, spectral types, proper motions and when available, Durchmusterung numbers and notes for the components of close to 100,000 systems based on ~600,000 means. The current version at the HEASARC is updated weekly and is derived from the version available online at https://crf.usno.navy.mil/wds/ (and mirrored at http://www.astro.gsu.edu/wds/), the latter being potentially updated nightly. The Washington Visual Double Star Catalog (WDS) is the successor to the Index Catalogue of Visual Double Stars, 1961.0 (IDS; Jeffers & van den Bos, 1963). Three earlier double star catalogs in the 20th century, those by Burnham (BDS; 1906), Innes (SDS; 1927), and Aitken (ADS; 1932), each covered only a portion of the sky. Both the IDS and the WDS cover the entire sky, and the WDS is intended to contain all known visual double stars for which at least one differential measure has been published. The WDS is continually updated as published data become available. Prior to this, two major updates have been published (Worley & Douglass 1984, 1997). The Washington Double Star Catalog (WDS) has seen numerous changes since the last major release of the catalog. The application of many techniques and considerable industry over the past few years has yielded unprecedented gains in both the number of systems and the number of measures. This version of the WDS catalog was first created at the HEASARC in March 2002 based on the USNO online version (available at either https://crf.usno.navy.mil/wds/ or http://www.astro.gsu.edu/wds/), and is updated by the HEASARC on at least a weekly basis. The table schema was last revised in February 2005. This is a service provided by NASA HEASARC .

This data set is part of a larger set of data called the Multibeam Bathymetry Database (MBBDB) where other similar data can be found

The TIGER/Line shapefiles and related database files (.dbf) are an extract of selected geographic and cartographic information from the U.S. Census Bureau's Master Address File / Topologically Integrated Geographic Encoding and Referencing (MAF/TIGER) Database (MTDB). The MTDB represents a seamless national file with no overlaps or gaps between parts, however, each TIGER/Line shapefile is designed to stand alone as an independent data set, or they can be combined to cover the entire nation. The Feature Names Relationship File (FEATNAMES.dbf) contains a record for each feature name and any attributes associated with it. Each feature name can be linked to the corresponding edges that make up that feature in the All Lines Shapefile (EDGES.shp), where applicable to the corresponding address range or ranges in the Address Ranges Relationship File (ADDR.dbf), or to both files. Although this file includes feature names for all linear features, not just road features, the primary purpose of this relationship file is to identify all street names associated with each address range. An edge can have several feature names; an address range located on an edge can be associated with one or any combination of the available feature names (an address range can be linked to multiple feature names). The address range is identified by the address range identifier (ARID) attribute, which can be used to link to the Address Ranges Relationship File (ADDR.dbf). The linear feature is identified by the linear feature identifier (LINEARID) attribute, which can be used to relate the address range back to the name attributes of the feature in the Feature Names Relationship File or to the feature record in the Primary Roads, Primary and Secondary Roads, or All Roads Shapefiles. The edge to which a feature name applies can be determined by linking the feature name record to the All Lines Shapefile (EDGES.shp) using the permanent edge identifier (TLID) attribute. The address range identifier(s) (ARID) for a specific linear feature can be found by using the linear feature identifier (LINEARID) from the Feature Names Relationship File (FEATNAMES.dbf) through the Address Range / Feature Name Relationship File (ADDRFN.dbf).