Try to figure out how to use SciKit Learn for logistic regression. Can you reproduce the results of the code you wrote?
Classify the molecules in the solubility dataset as "highly soluble" and "less soluble" depending on whether their measured solubility is above or below the median value for the data set. Use this data to train a logistic model for classifying molecules as highly or less soluble based on the molecular properties in the dataset.