CODEWITHSUNDEEP
pythonpandasnetworkx
user517696
user517696

Reputation: 2672

How to build a network from a dataframe

I have a dataframe like this:

User ID  Item   Category
U1       A      Furniture
U2       B      Sports
U3       C      Furniture
U2       A      Grocery
U3       B      Sports
U2.      B      Sports....

What I want to is to make a dictionary of users who have bought more than 3 same items as another user. For eg:

Lets say User U1 has bought Items A, B, C ,D, E, L, M. User U2 has bought Items A, B, C i.e 3 common items as User U1. User U3 bought B, C, L.

So if I want to find all such users who have bought atleast 3 items as U1, a dictionary should be returned in the following form

{U2: [A, B, C], U3:[B, C, L],....}

I have tried doing it with groupby() but it doesn't work. How do I achieve this??

Thanks

Upvotes: 0

Views: 68

Answers (1)

BENY
BENY

Reputation: 323226

IIUC

ID='U1'
n=1
Ux=df.loc[df.UserID==ID,'Item'].tolist()
s=df.loc[df.Item.isin(Ux)&~df.UserID.isin([ID]),].groupby('UserID').Item.count()
s1=s[s>=n].index.tolist()

d=df.loc[df.UserID.isin(s1),].groupby('UserID').Item.apply(list).to_dict()


d
Out[156]: {'U3': ['C'], 'U4': ['A']}

Upvotes: 3

Related Questions