Datasets
Title
Description
Filter(s)
1,300 Residential Smart-Meter Profiles
1,300 anonymized quarter-hour residential electricity profiles for 2023. |
Planet & Energy, Predictive Modelling & Forecasting
298 Residential Smart-Meter Time Series
298 anonymized half-hour residential smart-meter time series for January and February 2023. |
Planet & Energy, Predictive Modelling & Forecasting
Air Quality Observations for Belgium
Open air-quality observations reported to European authorities and structured for FAIR research; extendable to broader European data. |
Planet & Energy, Predictive Modelling & Forecasting
Anomaly Detection on Dynamic Knowledge Graphs
Synthetic dataset supporting anomaly-detection research on dynamic knowledge graphs. |
Industry, Anomaly Detection & Monitoring
bBSARD
Belgian legal-information dataset containing 22,417 French and Dutch statutory articles and about 1,100 legal questions linked to relevant articles. |
Society, Speech & Language AI
BEIR-NL
Dutch-language information-retrieval benchmark collection. The source inventory describes it as synthetically translated. |
Society, Speech & Language AI
CookDial
Task-oriented dialogue dataset with 260 human-to-human Wizard-of-Oz dialogues grounded in 260 cooking recipes. |
Society, Speech & Language AI
Cross-domain HTC
Mixed-domain dataset and code resource for cross-domain hierarchical text classification. |
Speech & Language AI, Society
Dialectloket Recordings
Recorded interviews with more than 300 speakers from the 1960s and 1970s, used as challenging Belgian Dutch speech archive material. |
Society, Speech & Language AI
EduQG
Dataset with 3,397 educational samples comprising multiple-choice questions, answers including distractors, and source documents. |
Society, Speech & Language AI
EmoTwiCS Customer Interactions
Real-world customer-interaction dataset for emotion and conversational-language research. |
Society, Speech & Language AI
EmoWOZ-CS EN-NL
English-Dutch emotional Wizard-of-Oz conversational dataset. |
Society, Speech & Language AI
European Geospatial Data for the Green Deal
European geospatial datasets referenced for FAIR environmental research and AI model training. |
Planet & Energy, Computer Vision
Humanity's Last Exam
Broad real-world benchmark dataset for evaluating multimodal AI systems. |
Society, Predictive Modelling & Forecasting
INSPIRE Datasets - EU & EFTA
European and EFTA geospatial datasets referenced for FAIR environmental research and AI model training. |
Planet & Energy, Computer Vision
JobHop
Large-scale public dataset of career trajectories extracted in a privacy-preserving manner from the VDAB database of CVs. |
Society, Predictive Modelling & Forecasting
LLM Ideology Analysis Dataset
Evaluations of political figures by a diverse set of large language models, enabling analysis of model ideology. |
Society, Explainable & Trustworthy AI
MSNER
Multilingual spoken named-entity recognition corpus with silver and gold annotations for VoxPopuli recordings in Dutch, French, German and Spanish. |
Society, Speech & Language AI
Multi-Sensor Voice Command Dataset
Audio recordings from a wireless network of four sensors, capturing five English voice commands and generic utterances from 20 volunteers across multiple positions, totaling 5.9 hours per sensor. |
Society, Speech & Language AI
Occupation Coding Dataset
Real-world datasets for automatic occupation coding. |
Society, Speech & Language AI
Physical Therapy Data
Publicly available 3D depth-camera data of people performing common physical-therapy exercises, supporting repetition segmentation, exercise recognition and exercise-quality analysis. |
Health, Computer Vision
Pléiades Neo Full Archive and Tasking
High- to very-high-resolution remote-sensing data referenced for FAIR environmental research and AI model training. |
Planet & Energy, Computer Vision
Pretrained CrIBo Checkpoints
Pretrained ViT-S/16 and ViT-B/16 vision-transformer checkpoints produced for CrIBo and trained on ImageNet-1k and/or COCO. |
Society, Computer Vision
RapidAI4EO Dataset
Large Earth-observation training dataset for machine-learning applications, referenced for FAIR environmental research. |
Society, Computer Vision
Replication Data for Personalized Keyword Spotting
400 speech samples recorded with an ultra-low-power microphone sensor after replaying a subset of HeySnips, with train/test splits, raw recordings and processed model outputs. |
Society, Speech & Language AI
RiGaD
Aerial imagery of rice seedlings for assessing germination rates and density. |
Planet & Energy, Computer Vision
SeizeIT1
EEG, behind-the-ear EEG and single-lead ECG recordings from 82 patients during presurgical epilepsy evaluations, with clinician-reviewed seizure annotations. |
Health, Anomaly Detection & Monitoring
Single-cell Open Problems Benchmark Data
Living benchmarking resource covering 12 single-cell analysis tasks and 81 datasets, used to evaluate methods with community-defined workflows and metrics. |
Health, Predictive Modelling & Forecasting
Water Observations for the Ijzer
Open observations from Waterinfo.be, structured and documented for use in FAIR environmental research. https://git.vito.be/projects/MARVIN-URBAN_WATER/repos/flandersai-water/browse |
Planet & Energy, Predictive Modelling & Forecasting