String "None" interpreted as NaN in newer versions of pandas #15
Description
The preprocessing code for chapter 12 doesn't work anymore. There are "None" strings for the holiday field of Metro_Interstate_Traffic_Volume.csv and newer versions of pandas interpret them as NaN. This can be fixed with the following modified call to pd_read_csv():
df = pd.read_csv('../data/Metro_Interstate_Traffic_Volume.csv', na_values=[], keep_default_na=False)