You can read about the bikesharing dataset here https://archive.ics.uci.edu/ml/datasets/bike+sharing+dataset. This dataset records the number of bikes rented in Capital bikeshare system in Washington D.C. and many other weather and environmental attributes.

Download the dataset. Using the daily data from day.csv file, create a label based on the attribute cnt into three categories {low, medium, high} demand based on some threshold values according to your judgment. Your task will be to create a nearest neighbor classifier for such a dataset using scikit-learn and evaluate your classifiers performance.

In your report, briefly describe on how you approached the problem, what interesting results you obtained, what practical difficulties you faced, and how you overcame these difficulties.