This folder contains the web application of the fraud prediciton model. We read data points lively from http://galvanize-case-study-on-fraud.herokuapp.com/data_point, which generate purchase record for prediction.

The system requirement:

• Python 2.7
• Pymongo 3.4.0
• flask 0.12.2
• cPickle 1.71
• urllib2 2.7

The record is read and made fraud prediciton based on a Random Forest Model we trained on preprietary data sets from Galvanize. User has three options to make prediction:

• option 1. Look at the fraudulent events prediction in the database.
• option 2. Read one entry from Galvanize website and make prediction
• option 3. User upload an entry from web app and make prediction

To run the app code.

python app.py