https://www.kaggle.com/c/talkingdata-adtracking-fraud-detection/data

understand the target and complete variables before you start

mini project above

1) use any language(python, sas, r) to provide data structures for all 4datasets but training is most important

2) import or read datasets into the environment

3) Detect missing values for all numeric variables

#missing %missing min max mean N =>table

4) Manipulate categorical variables (dummy, WOE or any other methods)

5) Treat missing values (any method)

6) provide complete data .sasdata ->.csv

7) provide data structure for complete data.

sas program(codes used in the program)->temp output ->data(csv or sas dataset) ->output data structure

the sas program and output structure should be in word files and output indicated that it is the final output

ods output needed in word or powerpoint

On monday

improve some features

do correlation study

feature engineering

consider prediction method

run model

submit result