Dataset Hub

This dataset is primarily used for analysis of energy and weather data, containing multiple features such as energy performance (Leistung) and weather data (cloud cover, radiation, temperature, etc.) from several geographic locations. The data exhibit various time lags and rolling statistics, making them suitable for time series analysis and forecasting models. The dataset is split into training, validation, and test sets with 61,368, 8,759, and 2,925 samples respectively.

The TreeSatAI-Time-Series dataset is a multi‑sensor collection for tree species classification tasks in the Central European region, based on forest management data from Lower Saxony, Germany, and includes labels for 20 European tree species. This dataset extends the original TreeSatAI dataset by integrating all available Sentinel‑1 and Sentinel‑2 time‑series data within a year to aid in distinguishing tree species.

The Weather dataset is a time‑series collection containing hourly weather condition information for a specific location. It records temperature, dew‑point temperature, relative humidity, wind speed, visibility, atmospheric pressure, and weather condition.

tiny-N-CMAPSS is a small‑scale version of the 2021 PHM Data Challenge dataset, a subset of the N-CMAPSS dataset. Sensor readings in the dataset have been converted from double‑precision to half‑precision floating‑point numbers, and the time‑series sampling rate has been reduced by a factor of 100 (from 1 Hz to 0.01 Hz). The training set consists of multiple files, while the test set is the competition’s test/validation set.

This dataset provides pre‑generated Mackey‑Glass time‑series numpy arrays generated with the jitcdde library. Because the library uses low‑level solvers, results can differ across machines even with identical ISA and library versions, so the pre‑generated data are recommended for reproducibility. Fourteen series are included; each uses MG parameters β = 0.2, γ = 0.1, n = 10 and τ ranging from 17 to 30. Each series spans 50 Lyapunov times, with 75 points per Lyapunov time.