This registry exists to help people discover and share datasets that are available via AWS resources. Learn more about sharing data on AWS.
If you want to add a dataset or example of how to use a dataset to this registry, please follow the instructions on the Registry of Open Data on AWS GitHub repository.
Unless specifically stated in the applicable dataset documentation, datasets available through the Registry of Open Data on AWS are not provided and maintained by AWS. Datasets are provided and maintained by a variety of third parties under a variety of licenses. Please check dataset licenses and related documentation to determine if a dataset may be used for your application.
The Sentinel-2 mission is a land monitoring constellation of two satellites that provide high resolution optical imagery and provide continuity for the current SPOT and Landsat missions. The mission provides a global coverage of the Earth's land surface every 5 days, making the data of great use in on-going studies. L1C data are available from June 2015 globally. L2A data are available from April 2017 over wider Europe region and globally since December 2018.
An ongoing collection of satellite imagery of all land on Earth produced by the Landsat 8 satellite.
A corpus of web crawl data composed of over 25 billion web pages.
This project creates a S3 repository with imagery acquired by the China-Brazil Earth Resources Satellite (CBERS). The image files are recorded and processed by Instituto Nacional de Pesquisa Espaciais (INPE) and are converted to Cloud Optimized Geotiff format in order to optimize its use for cloud based applications. The repository contains all CBERS-4 MUX, AWFI, PAN5M and PAN10M scenes acquired since the start of the satellite mission and is daily updated with new scenes.
Machine-readable data from certain electronic 990 forms filed with the IRS from 2013 to present.
A global dataset providing bare-earth terrain heights, tiled for easy usage and provided on S3.
This bucket contains multiple datasets (as Quilt packages) created by the Allen Institute for Cell Science (AICS). The imaging data in this bucket contains either of the following:1) field of view images from glass plates 2) cell membrane, DNA, and structure segmentations 3) cell membrane, DNA and structure contours 4) machine learning imaging predictions of the previously listed modalities.In addition, many of the datasets include CSVs that contain feature sets related to that data.
Real-time and archival data from the Next Generation Weather Radar (NEXRAD) network.
SpaceNet, launched in August 2016 as an open innovation project offering a repository of freely available imagery with co-registered map features. Before SpaceNet, computer vision researchers had minimal options to obtain free, precision-labeled, and high-resolution satellite imagery. Today, SpaceNet hosts datasets developed by its own team, along with data sets from projects like IARPA’s Functional Map of the World (fMoW).
The eBird Status and Trends project generates estimates of bird occurrence and abundance at a high spatiotemporal resolution. This dataset represents the primary modeled results from the analysis workflow and are designed for further analysis, synthesis, visualization, and exploration.
This project monitors the world's broadcast, print, and web news from nearly every corner of every country in over 100 languages and identifies the people, locations, organizations, counts, themes, sources, emotions, quotes, images and events driving our global society every second of every day.
This bucket contains multiple neuroimaging datasets (as Neuroglancer Precomputed Volumes) across multiple modalities and scales, ranging from nanoscale (electron microscopy), to microscale (cleared lightsheet microscopy and array tomography), and mesoscale (structural and functional magnetic resonance imaging). Additionally, many of the datasets include segmentations and meshes.
Amazon Customer Reviews (a.k.a. Product Reviews) is one of Amazon’s iconic products. In a period of over two decades since the first review in 1995, millions of Amazon customers have contributed over a hundred million reviews to express opinions and describe their experiences regarding products on the Amazon.com website. Over 130+ million customer reviews are available to researchers as part of this dataset.
The Community Earth System Model (CESM) Large Ensemble Numerical Simulation (LENS) dataset includes a 40-member ensemble of climate simulations for the period 1920-2100 using historical data (1920-2005) or assuming the RCP8.5 greenhouse gas concentration scenario (2006-2100), as well as longer control runs based on pre-industrial conditions. The data comprise both surface (2D) and volumetric (3D) variables in the atmosphere, ocean, land, and ice domains. The total data volume of the original dataset is ~500TB, which has traditionally been stored as ~150,000 individual CF/NetCDF files on disk or magnetic tape made available through the NCAR Climate Data Gateway for download or via web services. NCAR has copied a subset (currently ~70 TB) of CESM LENS data to Amazon S3 as part of the AWS Public Datasets Program. To optimize for large-scale analytics we have represented the data as ~275 Zarr stores format accessible through the Python Xarray library. Each Zarr store contains a single physical variable for a given model run type and temporal frequency (monthly, daily, 6-hourly).
Data of trips taken by taxis and for-hire vehicles in New York City.
OSM is a free, editable map of the world, created and maintained by volunteers. Regular OSM data archives are made available in Amazon S3.
Sentinel-1 is a pair of European radar imaging (SAR) satellites launched in 2014 and 2016. Its 6 days revisit cycle and ability to observe through clouds makes it perfect for sea and land monitoring, emergency response due to environmental disasters, and economic applications. GRD data is available globally since January 2017.
The Deutsche Börse Public Data Set consists of trade data aggregated to one minute intervals from the Eurex and Xetra trading systems. It provides the initial price, lowest price, highest price, final price and volume for every minute of the trading day, and for every tradeable security. If you need higher resolution data, including untraded price movements, please refer to our historical market data product here. Also, be sure to check out our developer's portal.
The Encyclopedia of DNA Elements (ENCODE) Consortium is an international collaboration of research groups funded by the National Human Genome Research Institute (NHGRI). The goal of ENCODE is to build a comprehensive parts list of functional elements in the human genome, including elements that act at the protein and RNA levels, and regulatory elements that control cells and circumstances in which a gene is active. ENCODE investigators employ a variety of assays and methods to identify functional elements. The discovery and annotation of gene elements is accomplished primarily by sequencing a diverse range of RNA sources, comparative genomics, integrative bioinformatic methods, and human curation. Regulatory elements are typically investigated through DNA hypersensitivity assays, assays of DNA methylation, and immunoprecipitation (IP) of proteins that interact with DNA and RNA, i.e., modified histones, transcription factors, chromatin regulators, and RNA-binding proteins, followed by sequencing.
Input data for the GEOS-Chem Chemical Transport Model. Including the NASA/GMAO MERRA-2 and GEOS-FP meteorological products, the HEMCO emission inventories, and other small data such as model initial conditions.
Select products from the Moderate Resolution Imaging Spectroradiometer (MODIS) managed by the U.S. Geological Survey and NASA.
The Wind Integration National Dataset (WIND) is an update and expansion of the Eastern Wind Integration Data Set and Western Wind Integration Data Set. It supports the next generation of wind integration studies.
Global, aggregated physical air quality data from public data sources provided by government, research-grade and other sources. These awesome groups do the hard work of measuring these data and publicly sharing them, and our community makes them more universally-accessible to both humans and machines.
SILO is a database of Australian climate data from 1889 to the present. It provides continuous, daily time-step data products in ready-to-use formats for research and operational applications. Gridded SILO data in annual NetCDF format are on AWS. Point data are available from the SILO website.
This dataset contains ground motion velocity and acceleration seismic waveforms recorded by the Southern California Seismic Network (SCSN) and archived at the Southern California Earthquake Data Center (SCEDC).
The goal of the USGS 3D Elevation Program (3DEP) is to collect elevation data in the form of light detection and ranging (LiDAR) data over the conterminous United States, Hawaii, and the U.S. territories, with data acquired over an 8-year period. This dataset provides two realizations of the 3DEP point cloud data. The first resource is a public access organization provided in Entwine Point Tiles format, which a lossless, full-density, streamable octree based on LASzip (LAZ) encoding. The second resource is a Requester Pays of the same data in LAZ (Compressed LAS) format. Resource names in both buckets correspond to the USGS project names.
This dataset contains soil infrared spectral data and paired soil property reference measurements for georeferenced soil samples that were collected through the Africa Soil Information Service (AfSIS) project, which lasted from 2009 through 2018. In this release, we include data collected during Phase I (2009-2013.) Georeferenced samples were collected from 19 countries in Sub-Saharan African using a statistically sound sampling scheme, and their soil properties were analyzed using both conventional soil testing methods and spectral methods (infrared diffuse reflectance spectroscopy). The two types of data can be paired to form a training dataset for machine learning, such that certain soil properties can be well-predicted through less expensive spectral techniques.
The Amazon Bin Image Dataset contains over 500,000 images and metadata from bins of a pod in an operating Amazon Fulfillment Center. The bin images in this dataset are captured as robot units carry pods as part of normal Amazon Fulfillment Center operations.
Global and high-resolution regional atmospheric models from Météo-France.
ERA5 is the fifth generation of ECMWF atmospheric reanalyses of the global climate, and the first reanalysis produced as an operational service. It utilizes the best available observation data from satellites and in-situ stations, which are assimilated and processed using ECMWF's Integrated Forecast System (IFS) Cycle 41r2. The dataset provides all essential atmospheric meteorological parameters like, but not limited to, air temperature, pressure and wind at different altitudes, along with surface parameters like rainfall, soil moisture content and sea parameters like sea-surface temperature and wave height. ERA5 provides data at a considerably higher spatial and temporal resolution than its legacy counterpart ERA-Interim. ERA5 consists of high resolution version with 31 km horizontal resolution, and a reduced resolution ensemble version with 10 members. It is currently available since 2008, but will be continuously extended backwards, first until 1979 and then to 1950. Learn more about ERA5 in Jon Olauson's paper ERA5: The new champion of wind power modelling?.
GOES satellites (GOES-16 & GOES-17) provide continuous weather imagery and monitoring of meteorological and space environment data across North America. GOES satellites provide the kind of continuous monitoring necessary for intensive data analysis. They hover continuously over one position on the surface. The satellites orbit high enough to allow for a full-disc view of the Earth. Because they stay above a fixed spot on the surface, they provide a constant vigil for the atmospheric "triggers" for severe weather conditions such as tornadoes, flash floods, hailstorms, and hurricanes. When these conditions develop, the GOES satellites are able to monitor storm development and track their movements.
The Hubble Space Telescope (HST) is one of the most productive scientific instruments ever created. This dataset contains calibrated and raw data for all of the currently active instruments on HST: ACS, COS, STIS and WFC3.
The National Agriculture Imagery Program (NAIP) acquires aerial imagery during the agricultural growing seasons in the continental U.S. This "leaf-on" imagery andtypically ranges from 60 centimeters to 100 centimeters in resolution and is available from the naip-analytic Amazon S3 bucket as 4-band (RGB + NIR) imagery in MRF format, on naip-source Amazon S3 bucket as 4-band (RGB + NIR) in uncompressed Raw GeoTiff format and naip-visualization as 3-band (RGB) Cloud Optimized GeotTiff format. NAIP data is delivered at the state level; every year, a number of states receive updates, with an overall update cycle of two or three years. More details on NAIP
A collection of Earth science datasets maintained by NASA, including climate change projections and satellite images of the Earth's surface.
The National Solar Radiation Database (NSRDB) is a serially complete collection of hourly and half-hourly values of the three most common measurements of solar radiation – global horizontal, direct normal, and diffuse horizontal irradiance — and meteorological data. These data have been collected at a sufficient number of locations and temporal and spatial scales to accurately represent regional solar radiation climates.
Open City Model is an initiative to provide cityGML data for all the buildings in the United States. By using other open datasets in conjunction with our own code and algorithms it is our goal to provide 3D geometries for every US building.
Grillo has developed an IoT-based earthquake early-warning system, with sensors currently deployed in Mexico, Chile and Costa Rica, and is now opening its entire archive of unprocessed accelerometer data to the world to encourage the development of new algorithms capable of rapidly detecting and characterizing earthquakes in real time.
Terra is the flagship satellite of NASA’s Earth Observing System (EOS). It was launched into orbit on December 18, 1999 and carries five instruments. These are the Moderate-resolution Imaging Spectroradiometer (MODIS), the Multi-angle Imaging SpectroRadiometer (MISR), the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), the Clouds and Earth’s Radiant Energy System (CERES), and the Measurements of Pollution in the Troposphere (MOPITT). The Terra Basic Fusion dataset is an easy-to-access record of the changes observed by all the instruments on the Terra mission for selected areas (described below) covering the years 2000-2015.
Meteorological data reusers now have an exciting opportunity to sample, experiment and evaluate Met Office atmospheric model data, whilst also experiencing a transformative method of requesting data via Restful APIs on AWS. For information about the data see the Met Office website. For examples of using the data check out the examples repository. If you need help and support using the data please raise an issue on the examples repository.Please note: Met Office continuously improves and updates its operational forecast models. Our last update became effective 04/12/2019. Please find the details here.
The 1000 Genomes Project is an international collaboration which has established the most detailed catalogue of human genetic variation, including SNPs, structural variants, and their haplotype context. The final phase of the project sequenced more than 2500 individuals from 26 different populations around the world and produced an integrated set of phased haplotypes with more than 80 million variants for these individuals.
Common reference genomes hosted on AWS S3. Can be used when aligning and analysing raw DNA sequencing data.
The Allen Brain Observatory – Visual Coding is a large-scale, standardized survey of physiological activity across the mouse visual cortex, hippocampus, and thalamus. It includes datasets collected with both two-photon imaging and Neuropixels probes, two complementary techniques for measuring the activity of neurons in vivo. The two-photon imaging dataset features visually evoked calcium responses from GCaMP6-expressing neurons in a range of cortical layers, visual areas, and Cre lines. The Neuropixels dataset features spiking activity from distributed cortical and subcortical brain regions, collected under analogous conditions to the two-photon imaging experiments. We hope that experimentalists and modelers will use these comprehensive, open datasets as a testbed for theories of visual information processing.
Broad maintained human genome reference builds hg19/hg38 and decoy references.
The Cell Painting Image Collection is a collection of freely downloadable microscopy image sets. Cell Painting is an unbiased high throughput imaging assay used to analyze perturbations in cell models. In addition to the images themselves, each set includes a description of the biological application and some type of "ground truth" (expected results). Researchers are encouraged to use these image sets as reference points when developing, testing, and publishing new image analysis algorithms for the life sciences. We hope that the this data set will lead to a better understanding of which methods are best for various biological image analysis applications.
Earth & Atmospheric Sciences at Cornell University has created a public data lake of climate data. The data is stored in columnar storage formats (ORC) to make it straightforward to query using standard tools like Amazon Athena or Apache Spark. The data itself is originally intended to be used for building decision support tools for farmers and digital agriculture. The first dataset is the historical NDFD / NDGD data distributed by NCEP / NOAA / NWS. The NDFD (National Digital Forecast Database) and NDGD (National Digital Guidance Database) contain gridded forecasts and observations at 2.5km resolution for the Contiguous United States (CONUS). There are also 5km grids for several smaller US regions and non-continguous territories, such as Hawaii, Guam, Puerto Rico and Alaska. NOAA distributes archives of the NDFD/NDGD via its NOAA Operational Model Archive and Distribution System (NOMADS) in Grib2 format. The data has been converted to ORC to optimize storage space and to, more importantly, simplify data access via standard data analytics tools.
The GATK test data resource bundle is a collection of files for resequencing human genomic data with the Broad Institute's Genome Analysis Toolkit (GATK).
These data correspond to the International LOFAR Telescope observations of the sky field ELAIS-N1 (16:10:01 +54:30:36) during the cycle 2 of observations. There are 11 runs of about 8 hours each plus the corresponding observation of the calibration targets before and after the target field. The data are measurement sets (MS) containing the cross-correlated data and metadata divided in 371 frequency sub-bands per target centred at ~150 MHz.
Global Historical Climatology Network - Daily is a dataset from NOAA that contains daily observations over global land areas. It contains station-based measurements from land-based stations worldwide, about two thirds of which are for precipitation measurement only. Other meteorological elements include, but are not limited to, daily maximum and minimum temperature, temperature at the time of observation, snowfall and snow depth. It is a composite of climate records from numerous sources that were merged together and subjected to a common suite of quality assurance reviews. Some data are more than 175 years old. The data is in CSV format. Each file corresponds to a year from 1763 to present and is named as such.
QIIME 2 is a powerful, extensible, and decentralized microbiome analysis package with a focus on data and analysis transparency. QIIME 2 enables researchers to start an analysis with raw DNA sequence data and finish with publication-quality figures and statistical results. This dataset contains the user docs (and related datasets) for QIIME 2.
Air Quality is a global SILAM atmospheric composition and air quality forecast performed on a daily basis for > 100 species and covering the troposphere and the stratosphere. The output produces 3D concentration fields and aerosol optical thickness. The data are unique: 20km resolution for global AQ models is unseen worldwide.
An ongoing collection of radiation and air quality measurements taken by devices involved in the Safecast project.
The S1 Single Look Complex (SLC) dataset contains Synthetic Aperture Radar (SAR) data in the C-Band wavelength. The SAR sensors are installed on a two-satellite (Sentinel-1A and Sentinel-1B) constellation orbiting the Earth with a combined revisit time of six days, operated by the European Space Agency. The S1 SLC data are a Level-1 product that collects radar amplitude and phase information in all-weather, day or night conditions, which is ideal for studying natural hazards and emergency response, land applications, oil spill monitoring, sea-ice conditions, and associated climate change effects.
The Cancer Genome Atlas (TCGA) is a joint effort of the National Cancer Institute (NCI) and the National Human Genome Research Institute (NHGRI) to accelerate our understanding of the molecular basis of cancer. TCGA-funded researchers across the United States have produced a corpus of raw and processed genomic, transcriptomic, and epigenomic data from thousands of cancer patients.
The Transiting Exoplanet Survey Satellite (TESS) is a two-year survey that will discover exoplanets in orbit around bright stars. More information about TESS is available at MAST and the TESS Science Support Center.
U.S. Census Bureau American Community Survey (ACS) Public Use Microdata Sample (PUMS) available in a linked data format using the Resource Description Framework (RDF) data model.
The Global Forecast System (GFS) is a weather forecast model produced by the National Centers for Environmental Prediction (NCEP). Dozens of atmospheric and land-soil variables are available through this dataset, from temperatures, winds, and precipitation to soil moisture and atmospheric ozone concentration. The entire globe is covered by the GFS at a base horizontal resolution of 18 miles (28 kilometers) between grid points, which is used by the operational forecasters who predict weather out to 16 days in the future. Horizontal resolution drops to 44 miles (70 kilometers) between grid point for forecasts between one week and two weeks.
VOiCES is a speech corpus recorded in acoustically challenging settings, using distant microphone recording. Speech was recorded in real rooms with various acoustic features (reverb, echo, HVAC systems, outside noise, etc.). Adversarial noise, either television, music, or babble, was concurrently played with clean speech. Data was recorded using multiple microphones strategically placed throughout the room. The corpus includes audio recordings, orthographic transcriptions, and speaker labels.
Uncompressed video used for video compression and video processing research.
This project primarily aims to facilitate performance benchmarking in robotics research. The dataset provides mesh models, RGB, RGB-D and point cloud images of over 80 objects. The physical objects are also available via the YCB benchmarking project. The data are collected by two state of the art systems: UC Berkley's scanning rig and the Google scanner. The UC Berkley's scanning rig data provide meshes generated with Poisson reconstruction, meshes generated with volumetric range image integration, textured versions of both meshes, Kinbody files for using the meshes with OpenRAVE, 600 High-resolution RGB images, 600 RGB-D images, and 600 point cloud images for each object. The Google scanner data provides 3 meshes with different resolutions (16k, 64k, and 512k polygons), textured versions of each mesh, Kinbody files for using the meshes with OpenRAVE.
3D models for molecular docking screens
The 3000 Rice Genome Project is an international effort to sequence the genomes of 3,024 rice varieties from 89 countries.
This dataset is the result of a collaborative project between the Communications Security Establishment (CSE) and The Canadian Institute for Cybersecurity (CIC) that use the notion of profiles to generate cybersecurity dataset in a systematic manner. It incluides a detailed description of intrusions along with abstract distribution models for applications, protocols, or lower level network entities. The dataset includes seven different attack scenarios, namely Brute-force, Heartbleed, Botnet, DoS, DDoS, Web attacks, and infiltration of the network from inside. The attacking infrastructure includes 50 machines and the victim organization has 5 departments includes 420 PCs and 30 servers. This dataset includes the network traffic and log files of each machine from the victim side, along with 80 network traffic features extracted from captured traffic using CICFlowMeter-V3. For more information on the creation of this dataset, see this paper by researchers at the Canadian Institute for Cybersecurity (CIC) and the University of New Brunswick (UNB): Toward Generating a New Intrusion Detection Dataset and Intrusion Traffic Characterization.
An open multi-sensor dataset for autonomous driving research. This dataset comprises semantically segmented images, semantic point clouds, and 3D bounding boxes. In addition, it contains unlabelled 360 degree camera images, lidar, and bus data for three sequences. We hope this dataset will further facilitate active research and development in AI, computer vision, and robotics for autonomous driving.
High resolution climate data to help assess the impacts of climate change primarily on agriculture. These open access datasets of climate projections will help researchers make climate change impact assessments.
COCO is a large-scale object detection, segmentation, and captioning dataset. This is part of the fast.ai datasets collection hosted by AWS for convenience of fast.ai students. If you use this dataset in your research please cite arXiv:1405.0312 [cs.CV].
LiDAR point cloud data for Washington, DC is available for anyone to use on Amazon S3. This dataset, managed by the Office of the Chief Technology Officer (OCTO), through the direction of the District of Columbia GIS program, contains tiled point cloud data for the entire District along with associated metadata.
This dataset contains historical and projected dynamically downscaled climate data for the State of Alaska and surrounding regions at 20km spatial resolution and hourly temporal resolution. Select variables are also summarized into daily resolutions. This data was produced using the Weather Research and Forecasting (WRF) model (Version 3.5). We downscaled both ERA-Interim historical reanalysis data (1979-2015) and both historical and projected runs from 2 GCM’s from the Coupled Model Inter-comparison Project 5 (CMIP5): GFDL-CM3 and NCAR-CCSM4 (historical run: 1970-2005 and RCP 8.5: 2006-2100).
Detailed air model results from EPA’s Risk-Screening Environmental Indicators (RSEI) model.
The data are from observations with the Murchison Widefield Array (MWA) which is a Square Kilometer Array (SKA) precursor in Western Australia. This particular dataset is from the Epoch of Reionization project which is a key science driver of the SKA. Nearly 2PB of such observations have been recorded to date, this is a small subset of that which has been exported from the MWA data archive in Perth and made available to the public on AWS. The data were taken to detect signatures of the first stars and galaxies forming and the effect of these early stars and galaxies on the evolution of the universe.
The Genome Ark hosts genomic information for the Vertebrate Genomes Project (VGP) and other related projects. The VGP is an international collaboration that aims to generate complete and near error-free reference genomes for all extant vertebrate species. These genomes will be used to address fundamental questions in biology and disease, to identify species most genetically at risk for extinction, and to preserve genetic information of life.
Several reference genomes to enable translation of whole human genome sequencing to clinical practice.
GSOD is a collection of daily weather measurements (temperature, wind speed, humidity, pressure, and more) from 9000+ weather stations around the world.
N-grams are fixed size tuples of items. In this case the items are words extracted from the Google Books corpus. The n specifies the number of elements in the tuple, so a 5-gram contains five words or characters. The n-grams in this dataset were produced by passing a sliding window of the text of books and outputting a record for each new token.
HIRLAM (High Resolution Limited Area Model) is an operational synoptic and mesoscale weather prediction model managed by the Finnish Meteorological Institute.
Population data for a selection of countries, allocated to 1 arcsecond blocks and provided in a combination of CSV and Cloud-optimized GeoTIFF files. This refines CIESIN’s Gridded Population of the World using machine learning models on high-resolution worldwide Digital Globe satellite imagery. CIESIN population counts aggregated from worldwide census data are allocated to blocks where imagery appears to contain buildings.
This dataset includes sequencing data, assemblies, and analyses for the offspring of ten parent-offspring trios.
The International Cancer Genome Consortium (ICGC) coordinates projects with the common aim of accelerating research into the causes and control of cancer. The PanCancer Analysis of Whole Genomes (PCAWG) study is an international collaboration to identify common patterns of mutation in whole genomes from ICGC. More than 2,400 consistently analyzed genomes corresponding to over 1,100 unique ICGC donors are now freely available on Amazon S3 to credentialed researchers subject to ICGC data sharing policies.
The IChangeMyCity project provides insight into the complaints raised by citizens from diffent cities of India related to the issues in their neighbourhoods and the resolution of the same by the civic bodies.
Excerpts of electronic Form 990 and 990-EZ filings, converted to spreadsheet form. Additional fields being added regularly.
Some of the most important datasets for image classification research, including CIFAR 10 and 100, Caltech 101, MNIST, Food-101, Oxford-102-Flowers, Oxford-IIIT-Pets, and Stanford-Cars. This is part of the fast.ai datasets collection hosted by AWS for convenience of fast.ai students. See documentation link for citation and license details for each dataset.
Some of the most important datasets for image localization research, including Camvid and PASCAL VOC (2007 and 2012). This is part of the fast.ai datasets collection hosted by AWS for convenience of fast.ai students. See documentation link for citation and license details for each dataset.
Dataset and benchmarks for computer vision research in the context of autonomous driving. The dataset has been recorded in and around the city of Karlsruhe, Germany using the mobile platform AnnieWay (VW station wagon) which has been equipped with several RGB and monochrome cameras, a Velodyne HDL 64 laser scanner as well as an accurate RTK corrected GPS/IMU localization unit. The dataset has been created for computer vision and machine learning research on stereo, optical flow, visual odometry, semantic segmentation, semantic instance segmentation, road segmentation, single image depth prediction, depth map completion, 2D and 3D object detection and object tracking. In addition, several raw data recordings are provided. The datasets are captured by driving around the mid-size city of Karlsruhe, in rural areas and on highways. Up to 15 cars and 30 pedestrians are visible per image.
The Kepler mission observed the brightness of more than 180,000 stars near the Cygnus constellation at a 30 minute cadence for 4 years in order to find transiting exoplanets, study variable stars, and find eclipsing binaries. More information about the Kepler mission is available at MAST.
The Multimedia Commons is a collection of audio and visual features computed for the nearly 100 million Creative Commons-licensed Flickr images and videos in the YFCC100M dataset from Yahoo! Labs, along with ground-truth annotations for selected subsets. The International Computer Science Institute (ICSI) and Lawrence Livermore National Laboratory are producing and distributing a core set of derived feature sets and annotations as part of an effort to enable large-scale video search capabilities. They have released this feature corpus into the public domain, under Creative Commons License 0, so it is free for anyone to use for any purpose.
Some of the most important datasets for NLP, with a focus on classification, including IMDb, AG-News, Amazon Reviews (polarity and full), Yelp Reviews (polarity and full), Dbpedia, Sogou News (Pinyin), Yahoo Answers, Wikitext 2 and Wikitext 103, and ACL-2010 French-English 10^9 corpus. This is part of the fast.ai datasets collection hosted by AWS for convenience of fast.ai students. See documentation link for citation and license details for each dataset.
In order to support NOAA's homeland security and emergency response requirements, the National Geodetic Survey Remote Sensing Division (NGS/RSD) has the capability to acquire and rapidly disseminate a variety of spatially-referenced datasets to federal, state, and local government agencies, as well as the general public. Remote sensing technologies used for these projects have included lidar, high-resolution digital cameras, a film-based RC-30 aerial camera system, and hyperspectral imagers. Examples of rapid response initiatives include acquiring high resolution images with the Emerge/Applanix Digital Sensor System (DSS).
The Global Ensemble Forecast System (GEFS), previously known as the GFS Global ENSemble (GENS), is a weather forecast model made up of 21 separate forecasts, or ensemble members. The National Centers for Environmental Prediction (NCEP) started the GEFS to address the nature of uncertainty in weather observations, which is used to initialize weather forecast models. The GEFS attempts to quantify the amount of uncertainty in a forecast by generating an ensemble of multiple forecasts, each minutely different, or perturbed, from the original observations. With global coverage, GEFS is produced four times a day with weather forecasts going out to 16 days.
The Global Forecast System (GFS) is a weather forecast model produced by the National Centers for Environmental Prediction (NCEP). Dozens of atmospheric and land-soil variables are available through this dataset, from temperatures, winds, and precipitation to soil moisture and atmospheric ozone concentration. The entire globe is covered by the GFS at a base horizontal resolution of 18 miles (28 kilometers) between grid points, which is used by the operational forecasters who predict weather out to 16 days in the future. Horizontal resolution drops to 44 miles (70 kilometers) between grid point for forecasts between one week and two weeks. Both the current version and the FV3-based parallel version of the GFS being tested to become the new operational model at a future date are available.
Global Hydro-Estimator provides a global mosaic imagery of rainfall estimates from multi-geostationary satellites, which currently includes GOES-16, GOES-15, Meteosat-8, Meteosat-11 and Himawari-8. The GHE products include: Instantaneous rain rate, 1 hour, 3 hour, 6 hour, 24 hour and also multi-day rainfall accumulation.
The HRRR is a NOAA real-time 3-km resolution, hourly updated, cloud-resolving, convection-allowing atmospheric model, initialized by 3km grids with 3km radar assimilation. Radar data is assimilated in the HRRR every 15 min over a 1-h period adding further detail to that provided by the hourly data assimilation from the 13km radar-enhanced Rapid Refresh.
The Integrated Surface Database (ISD) consists of global hourly and synoptic observations compiled from numerous sources into a gzipped fixed width format. ISD was developed as a joint activity within Asheville's Federal Climate Complex. The database includes over 35,000 stations worldwide, with some having data as far back as 1901, though the data show a substantial increase in volume in the 1940s and again in the early 1970s. Currently, there are over 14,000 "active" stations updated daily in the database. The total uncompressed data volume is around 600 gigabytes; however, it continues to grow as more data are added. ISD includes numerous parameters such as wind speed and direction, wind gust, temperature, dew point, cloud data, sea level pressure, altimeter setting, station pressure, present weather, visibility, precipitation amounts for various time periods, snow depth, and various other elements as observed by each station.
The NOAA National Water Model Reanalysis dataset contains output from multi-decade retrospective simulations. These simulations used observed rainfall as input and ingested other required meteorological input fields from a weather reanalysis dataset. The output frequency and fields available in this historical NWM dataset differ from those contained in the real-time forecast model. One application of this dataset is to provide historical context to current real-time streamflow, soil moisture and snowpack NWM conditions. The reanalysis data can be used to infer flow frequencies and perform temporal analyses with hourly streamflow output and 3-hourly land surface output. The long-term dataset can also be used in the development of end user applications which require a long baseline of data for system training or verification purposes. This dataset contains output from two retrospective simulations. Currently there are two versions of the dataset: A 25-year (January 1993 through December 2017) retrospective simulation using version 1.2 of the National Water Model, and a 26-year (January 1993 through December 2018) retrospective simulation using version 2.0 of the National Water Model.
The National Water Model (NWM) is a water resources model that simulates and forecasts water budget variables, including snowpack, evapotranspiration, soil moisture and streamflow, over the entire continental United States (CONUS). The model, launched in August 2016, is designed to improve the ability of NOAA to meet the needs of its stakeholders (forecasters, emergency managers, reservoir operators, first responders, recreationists, farmers, barge operators, and ecosystem and floodplain managers) by providing expanded accuracy, detail, and frequency of water information. It is operated by NOAA’s Office of Water Prediction. This bucket contains a four-week rollover of the Short Range Forecast model output and the corresponding forcing data for the model. The model is forced with meteorological data from the High Resolution Rapid Refresh (HRRR) and the Rapid Refresh (RAP) models. The Short Range Forecast configuration cycles hourly and produces hourly deterministic forecasts of streamflow and hydrologic states out to 18 hours.
The Operational Forecast System (OFS) has been developed to serve the maritime user community. OFS was developed in a joint project of the NOAA/National Ocean Service (NOS)/Office of Coast Survey, the NOAA/NOS/Center for Operational Oceanographic Products and Services (CO-OPS), and the NOAA/National Weather Service (NWS)/National Centers for Environmental Prediction (NCEP) Central Operations (NCO). OFS generates water level, water current, water temperature, water salinity (except for the Great Lakes) and wind conditions nowcast and forecast guidance four times per day.
This dataset includes the sequencing and assembly of a reference standard human genome (GM12878) using the MinION nanopore sequencing instrument with the R9.4 1D chemistry.
A free software, global observation network for detecting censorship, surveillance and traffic manipulation on the internet.
OpenNeuro is a database of openly-available brain imaging data. The data are shared according to a Creative Commons CC0 license, providing a broad range of brain imaging data to researchers and citizen scientists alike. The database primarily focuses on functional magnetic resonance imaging (fMRI) data, but also includes other imaging modalities including structural and diffusion MRI, electroencephalography (EEG), and magnetoencephalograpy (MEG). OpenfMRI is a project of the Center for Reproducible Neuroscience at Stanford University. Development of the OpenNeuro resource has been funded by the National Science Foundation, National Institute of Mental Health, National Institute on Drug Abuse, and the Laura and John Arnold Foundation.
OSMLR a linear referencing system built on top of OpenStreetMap. OSM has great information about roads around the world and their interconnections, but it lacks the means to give a stable identifier to a stretch of roadway. OSMLR provides a stable set of numerical IDs for every 1 kilometer stretch of roadway around the world. In urban areas, OSMLR IDs are attached to each block of roadways between significant intersections.
PhysioNet offers free web access to large collections of recorded physiologic signals (PhysioBank) and related open-source software (PhysioToolkit).
ParaCrawl is a set of large parallel corpora to/from English for all official EU languages by a broad web crawling effort. State-of-the-art methods are applied for the entire processing chain from identifying web sites with translated text all the way to collecting, cleaning and delivering parallel corpora that are ready as training data for CEF.AT and translation memories for DG Translation.
Software Heritage is the largest existing public archive of software source code and accompanying development history. The Software Heritage Graph Dataset is a fully deduplicated Merkle DAG representation of the Software Heritage archive.The dataset links together file content identifiers, source code directories, Version Control System (VCS) commits tracking evolution over time, up to the full states of VCS repositories as observed by Software Heritage during periodic crawls. The dataset’s contents come from major development forges (including GitHub and GitLab), FOSS distributions (e.g., Debian), and language-specific package managers (e.g., PyPI). Crawling information is also included, providing timestamps about when and where all archived source code artifacts have been observed in the wild.
Tabula Muris is a compendium of single cell transcriptomic data from the model organism Mus musculus comprising more than 100,000 cells from 20 organs and tissues. These data represent a new resource for cell biology, reveal gene expression in poorly characterized cell populations, and allow for direct and controlled comparison of gene expression in cell types shared between tissues, such as T-lymphocytes and endothelial cells from different anatomical locations. Two distinct technical approaches were used for most organs: one approach, microfluidic droplet-based 3’-end counting, enabled the survey of thousands of cells at relatively low coverage, while the other, FACS-based full length transcript analysis, enabled characterization of cell types with high sensitivity and coverage. The cumulative data provide the foundation for an atlas of transcriptomic cell biology. See: https://www.nature.com/articles/s41586-018-0590-4
Tabula Muris Senis is a comprehensive compendium of single cell transcriptomic data from the model organism Mus musculus comprising more than 500,000 cells from 18 organs and tissues across the mouse lifespan. We discovered cell-specific changes occurring across multiple cell types and organs, as well as age related changes in the cellular composition of different organs. Using single-cell transcriptomic data we were able to assess cell type specific manifestations of different hallmarks of aging, such as senescence, changes in the activity of metabolic pathways, depletion of stem-cell populations, genomic instability and the role of inflammation as well as other changes in the organism’s immune system. Tabula Muris Senis provides a wealth of new molecular information about how the most significant hallmarks of aging are reflected in a broad range of tissues and cell types.See: https://www.biorxiv.org/content/10.1101/661728v1
The Genome Institute at Washington University has developed a high-throughput, fault-tolerant analysis information management system called the Genome Modeling System (GMS), capable of executing complex, interdependent, and automated genome analysis pipelines at a massive scale. The GMS framework provides detailed tracking of samples and data coupled with reliable and repeatable analysis pipelines. GMS includes a full system image with software and services, expandable from one workstation to a large compute cluster.
The Human Connectome Project aims to provide an unparalleled compilation of neural data, an interface to graphically navigate this data and the opportunity to achieve never before realized conclusions about the living human brain.
The NIH-funded Human Microbiome Project (HMP) is a collaborative effort of over 300 scientists from more than 80 organizations to comprehensively characterize the microbial communities inhabiting the human body and elucidate their role in human health and disease. To accomplish this task, microbial community samples were isolated from a cohort of 300 healthy adult human subjects at 18 specific sites within five regions of the body (oral cavity, airways, urogenital track, skin, and gut). Targeted sequencing of the 16S bacterial marker gene and/or whole metagenome shotgun sequencing was performed for thousands of these samples. In addition, whole genome sequences were generated for isolate strains collected from human body sites to act as reference organisms for analysis. Finally, 16S marker and whole metagenome sequencing was also done on additional samples from people suffering from several disease conditions.
MMID is a large-scale, massively multilingual dataset of images paired with the words they represent collected at the University of Pennsylvania. The dataset is doubly parallel: for each language, words are stored parallel to images that represent the word, and parallel to the word's translation into English (and corresponding images.)
GOES provides continuous weather imagery and monitoring of meteorological and space environment data across North America.
MIMIC-III (‘Medical Information Mart for Intensive Care’) is a large, single-center database comprising information relating to patients admitted to critical care units at a large tertiary care hospital. Data includes vital signs, medications, laboratory measurements, observations and notes charted by care providers, fluid balance, procedure codes, diagnostic codes, imaging reports, hospital length of stay, survival data, and more. The database supports applications including academic and industrial research, quality improvement initiatives, and higher education coursework. The MIMIC-III dataset is freely-available. Researchers seeking to use the database must formally request access. For details, see the getting started page. Once you have a PhysioNet account, you must enable acccess to the MIMIC-III dataset from your AWS account. To do this, please input your AWS account number, and request access to the MIMIC-III Clinical Database on AWS.
Data from the Moderate Resolution Imaging Spectroradiometer (MODIS), managed by the U.S. Geological Survey and NASA. Five products are included: MCD43A4 (MODIS/Terra and Aqua Nadir BRDF-Adjusted Reflectance Daily L3 Global 500 m SIN Grid), MOD11A1 (MODIS/Terra Land Surface Temperature/Emissivity Daily L3 Global 1 km SIN Grid), MYD11A1 (MODIS/Aqua Land Surface Temperature/Emissivity Daily L3 Global 1 km SIN Grid), MOD13A1 (MODIS/Terra Vegetation Indices 16-Day L3 Global 500 m SIN Grid), and MYD13A1 (MODIS/Aqua Vegetation Indices 16-Day L3 Global 500 m SIN Grid). MCD43A4 has global coverage, all time (~21 years). The other products have ~11 years of global coverage. All data files are in single-band cloud-optimized GeoTIFF (COG) format.
The basic geo-data set for public transport stops comprises public transport stops in Switzerland and additional selected geo-referenced public transport locations that are of operational or structural importance (operating points).
Pre and post event high-resolution satellite imagery in support of emergency planning, risk assessment, monitoring of staging areas and emergency response, damage assessment, and recovery. Also incudes crowdsourced damage assessments for major, sudden onset disasters.
The Multiview Extended Video with Activities (MEVA) dataset consists video data of human activity, both scripted and unscripted, collected with roughly 100 actors over several weeks. The data was collected with 29 cameras with overlapping and non-overlapping fields of view. The current release consists of about 405 gigabytes across 3370 video clips; the total video length is about 254 hours. Further updates are planned. Available metadata includes a site map and camera models for some of the cameras.