Loading...
Loading...
Text classification, translation, QA, summarization, dialogue, sentiment analysis, language modeling, text corpora
49,576 datasets
A national modelled dataset showing flood depths on a 20x20 metre grid across the UK. It was created using a Digital Terrain Model and represents a 1 in 1000 annual chance fluvial flood scenario, with depths filtered to the 0-10 metre range. The data is provided by the Environment Agency and incorporates spatial data from multiple UK government and research organizations.
Quarterly updated data from the UK Environment Agency's National Incident Recording System (NIRS2) details substantiated, completed, and closed environmental management incidents. The dataset is filtered to include only incidents with a Category 1 (major) or Category 2 (significant) impact level on at least one environmental medium (water, land, or air). Records are subject to an inherent investigation lag and may change due to final quality assurance.
Geoscience Australia produced seabed morphology and geomorphology maps for the Flinders Reefs and Cairns Seamount within the Coral Sea Marine Park. The maps were created using a two-step classification system applied to bathymetry digital elevation models supplemented by backscatter intensity, seabed imagery, and sediment samples. The data product is intended for marine park managers, regulators, and the general public.
A dataset from Princeton University modeling hot-electron-induced catalytic reactions. The data likely contains potential energy surfaces for hydrogen dissociation on Au(111) surfaces, calculated using embedded correlated wavefunction theory. It is authored by Florian Libisch and sourced from the paperswithcode platform.
Farzin Javanshour of Tampere University reports the effect of a 1.2 wt% graphene oxide coating on the tensile behavior of single flax fibre bundles. The dataset includes results for four treatment subclasses: as-received fibres, fibres soaked in water for 30 minutes, fibres soaked for 24 hours, and fibres treated with graphene oxide for 24 hours. Tensile properties were measured using the impregnated fibre bundle test method with an Instron 5567 machine and a 30 kN loadcell.
STARSS23 contains multichannel recordings of sound scenes in various rooms and environments, with temporal and spatial annotations of prominent sound events. The dataset was collected by Tampere University in Finland and SONY in Japan between September 2021 and February 2023, serving as the development dataset for the DCASE 2023 Challenge. It includes over 4 hours of additional recordings compared to its predecessor, featuring 360° videos and distance labels for sound events.
L0DROP, a stochastic gating layer, prunes 40-70% of Transformer encoder outputs without significant quality loss. The method was tested on two machine translation and two summarization tasks by Biao Zhang of the University of Edinburgh. Pruning yields decoding speedups of up to 1.65x on document summarization and shows systematic preferences for pruning function words and punctuation.
Friction data from ring-shear tests on an iron powder and quartz sand mixture with a 1:3 weight ratio. The material was characterized by the Helmholtz Laboratory for Tectonic Modelling at GFZ Potsdam in 2017, providing peak, dynamic, and reactivation friction coefficients and cohesion values. This dataset supports analog modelling experiments monitored with CT scanners at Utrecht University.
NOAA20 satellite data provides daily chlorophyll-a concentration estimates derived from VIIRS sensor multi-spectral measurements using the Garver-Siegel-Maritorena (GSM) algorithm. The data is processed by the SeaDAS software l2gen and described in IOCCG Report 5, 2006. It is provided by the Australian Ocean Data Network and was last updated on 2026-07-10.
Beagle Marine Park in south-eastern Australia contains geospatial data on seabed morphology and geomorphology. The data product was created using a nationally consistent two-part classification system, applying semi-automated GIS tools to bathymetry digital elevation models. It is intended for use by marine park managers, regulators, and the general public.
Port Phillip Bay, Australia, is the focus of this dataset on benthic nutrient recycling. It contains chamber measurements of oxygen, ammonium, nitrate, nitrite, phosphate, silicate, TCO2, and alkalinity from various sites during the summers of 1994 and 1995. The data, sourced from the Australian Ocean Data Network, defines solute exchange rates between sediment and water, quantifying nutrient regeneration and denitrification efficiency.
IMOS provides satellite-derived chlorophyll-a concentration data from the MODIS sensor on the Aqua platform, using the Carder retrieval algorithm. Production of contemporary MODIS data ceased at 2024-12-26, with users directed to VIIRS instruments for newer data. The dataset was last updated on the Australian Ocean Data Network platform in July 2026.
The Aqua satellite's MODIS sensor provides multi-spectral measurements of reflected sunlight from the ocean surface layer. These data use the Garver-Siegel-Maritorena (GSM) algorithm to infer chlorophyll-a concentration, primarily from phytoplankton. Production of contemporary MODIS data ceased on 2024-12-26, with users directed to VIIRS instruments for newer data.
Scotland's national forest estate is mapped through an interactive web service containing nine data layers. The layers include recreation points, routes, areas, roads, bridges, sub-compartments, ownership, legal boundaries, and the Galloway Dark Skies Park. Forestry Commission Scotland, with cooperation from Scottish Natural Heritage, provides this service, which was last updated on July 8, 2026.
20 London boroughs participate in a scheme recognizing caterers offering healthier food options. The London Borough of Barnet, in partnership with the Greater London Authority and NHS Barnet, runs this award program to encourage reduced saturated fat and salt. Restaurants, takeaways, sandwich bars, pubs, and other caterers can apply for recognition, which includes a certificate and window sticker.
MODIS sensor data from the Aqua satellite provides daily estimates of chlorophyll-a concentration in ocean surface waters using the OCI algorithm. The dataset is produced by the NASA Ocean Biology Processing Group and distributed via the Australian Ocean Data Network. Production of contemporary MODIS data ceased on 2024-12-26, with users directed to VIIRS instruments for newer data.
Swiss smart meter data from Canton Lucerne covers a two-year period from January 1, 2021, to December 31, 2022. It contains active energy consumption values in kWh recorded at 15-minute intervals for 4,959 individual metering points, all of which have complete data for the full duration. The dataset, cleaned and anonymized by Benjamin Sawicki of ETH Zurich, is based on a collection from CKW AG.
Distribution data for benthic genera in the North Sea, covering a 22-year timespan from 1990 to 2011. This data accompanies a manuscript in the Journal of Applied Ecology and was obtained from re-analysis of OBIS data supplemented with Cefas survey data, with the study region partitioned into a 0.25° x 0.25° regular grid.
IMOS provides satellite-derived ocean data from the NOAA20 platform's VIIRS sensor. The dataset contains daily estimates of the diffuse attenuation coefficient at 490nm wavelength (Kd490), which indicates light penetration in the water column. The algorithm is based on a semi-empirical model using water-leaving radiances at 490nm and 555nm, as described by NASA's Ocean Color team.
IMOS provides daily chlorophyll-a concentration data derived from the MODIS sensor on NASA's Aqua satellite using the OC3 algorithm. The dataset is produced by the Australian Ocean Data Network and was last updated on 2026-07-10. Production of contemporary MODIS data ceased at 2024-12-26, with users directed to VIIRS instruments for newer data.