sfdat28/giuli-project-ideas at master · giuperrotti/sfdat28 · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
Analyze Lending Club Notes Titles to predict default
	- an NLP approach to risk:
		- ARE FULLY CAPITALIZED TITLES DIFFERENT?
		- Does the number of word have an impact on default?
	- is the term a good predictor of churn?
	- is the loan intent a good predictor of churn?

Predict Approval Odds for PLoans and CCs
	- improve underwriting criteria for credit products
	- understand what are the relevant features
	- understand how to create a better and self-sustainable feedbac loop to improve the model

Predict who are the users likely to login / churn
	- given how users interact with the website
		- what is the page form factor of registration?
		- does the user have our app?
		- how many times have this user logged in the past?
			- past week
			- past month
			- past 6 months
	- given how users interact with email
		- how many emails did we send?
			- per email category
		- how many emails did the user open / click / engage with?
			- per email category
		- has the user unsubscribed?
			- per email category
	- given the user's features
		- what is the user's credit score?
		- what are the users demographics?
		- what is the marketing source?

Predict who are the users likely to unsubscribe
	- given how users interact with the website
		- what is the page form factor of registration?
		- does the user have our app?
		- how many times have this user logged in the past?
			- past week
			- past month
			- past 6 months
	- given how users interact with email
		- how many emails did we send?
			- per email category
		- how many emails did the user open / click / engage with?
			- per email category
		- has the user unsubscribed?
			- per email category
	- given the user's features
		- what is the user's credit score?
		- what are the users demographics?
		- what is the marketing source?