recycling_bin.py

'''
**I'll check for (and remove) outliers in IPs' network and host seperately.**
'''

df.network.value_counts().tail(100)

df = df.groupby('network').filter(lambda x : len(x)>3)
df.network.value_counts()

df.shape

df.host.value_counts().tail(800)

df = df.groupby('host').filter(lambda x : len(x)>2)
df.host.value_counts()

df.shape

'''
I might choose to explore hosts with only one-time access at a later point, in which case I would re-add these dropped observations.

I decided to take this code out because it removed too many anomalies that could be interesting. 

Saving this code here for documentation purposes, and in case I want to use it again later.
'''





for x in darden_df.ip: 
    try:
#        response = reader.city(x)
        darden_df['ip_city'] = reader.city(x).city.name
        break
    except:
        print("Oops!", sys.exc_info()[0], "occurred.")
        print("Next observation.")
        darden_df['ip_city'] = 'error'



for x in darden_df.ip: 
    try:
        response = reader.city(x)
        darden_df['ip_city'] = response.city.name
        break
    except:
        print("Oops!", sys.exc_info()[0], "occurred.")
        print("Next observation.")
        darden_df['ip_city'] = 'error'