QFE Project

This is the Python backend for my Computer Science 523 class (Software Engineering). The purpose of this application is to retrieve price information from Quandl on a set of financial assets, convert them to log returns, form a portfolio, and optimize that portfolio given a set of parameters. Overall, this code creates the necessary data for our frontend view

Getting Started

To see the system running, visit https://qfe-app-523-maguilar.cloudapps.unc.edu/#!/. There is an additional README file on this page with more specific instructions

If you want to install and run the backend on a local device, installing Anaconda is recommended. Once Anaconda is installed, add all dependancies using the method mentioned in "Prerequisites". Note, any interface attempting to access the locally running backend will need to update URLs for AJAX calls. Use Python 3's built in http.server method to run the backend.

Prerequisites

To install depenencies use "pip install -r requirements.txt" from the directory where the code is located. For a list of dependancies, check out requirements.txt

Layout

app.py = The entry point of the application. This is where all the routing is handled and calls to other modules are made. This is also the home of the flask server

benchmark.py = The object dedicated to forming and returning information about a user provided benchmark

portfolio.py = The object dedicated to forming and returning information about a set of user provided assets given a date range, rebalancing frequency, etc

data.py = The module used for data preprocessing needed in portfolio.py and benchmark.py

data_test_suite.py = All unit tests related to the data.py object

portfolio_test_suite.py = All unit tests related to the portfolio.py and benchmark.py objects

benchmark_test_suite.py = All unit tests related to the benchmark.py objects

Running the tests

Navigate to the directory containing any of the test suites. Then, for example, run "pytest portfolio_test_suite.py" and the unit tests for the selected module should execute successfully. An internet connection is required to run the tests. Also, the code must be on your local device, but it does not need to be served up at the same time (IE, using python's http.server).

Deployment

See the client handoff plan on our website for more details http://test-hoelzel.cloudapps.unc.edu/#!/home. The application is deployed on Carolina CloudApps under UNC Economics department credentials

Contributing

Nicholas McHenry

Douglas Hoelzel

Avery Lue

Kurtis Bass

Author