Synthetic time-series data that your algorithm should manage to forecast with a 100% accuracy
This is a collection of hand-crafted, synthetic time-series data (and R / Python scripts used to generate them). All of these data-sets exhibit a strong relationship between input and output (almost no noise). They can be used as a baseline for any forecasting / predictive analytics algorithm, to see if (and how fast) they manage to model and predict these patterns.