If you want to donate a dataset with concept drift from a real-world streaming problem, please contact Vinicius Souza (vinicius at ppgia.pucpr.br) or Gustavo Batista (g.batista at unsw.edu.au) answering the following questions:
Where can we get the data in ARFF/CSV format?
Is there a publication related to the data?
Can you describe the classification problem related to the data, and why is it a streaming problem?
Can you explain the order of the examples and the interval between observations?
Can you describe the class labels?
Do you know the type of drift (incremental, gradual, abrupt, reoccurring) which occur in the dataset and when they occur?