big-portfolio-learner v1.0.0

Final project for CIS 545 Big Data Analytics.

The project proposal can be found here

Prerequisites and setup

• requires an Alpha Vantage API key and knowledge of the API. For documentation of the Alpha Vantage API, please go (here)[https://www.alphavantage.co/documentation/]

• assumes the working directory has a file called secrets.yml which contains the credentials for the data APIs. For Alpha Vantage, there should be a line alpha_key: "[YOUR API KEY]".

• requirements.txt contains the packages and versions used thus far. Generated using pipreqs. More information on pipreqs can be found here.

• install packages using pip install -r requirements.txt or sudo -H pip install -r requirements.txt.

Analysis and Portfolio Learner

notebooks/*.ipynb

The subdirectory notebooks/ contains the iPython notebooks designed to run on Google CoLab.

These notebook pull the data from an S3 bucket, do some Exploratory Data Analysis, clean and munges the data, and finally implements the core logic of predicting which stocks to invest in using time series analysis.

NOTE: Due to the time constraints, we were only able to finish up to step3a_single_tsa.ipynb where we implement a time series analysis on a single stock. step3b_global_tsa.ipynb was intended to implement a global time series analysis on all stocks' time series, but wasn't finished.

Pulling data from the Alpha Vantage API

DataLoader.py

Contains the interface for loading data from Alpha Vantage API.

The DataLoader will call the API and write .csv files (for each stock ticker) to the output paths provided in the order of the stock ticker symbols passed to it.

The last step is to zip up the data directories so that it can be manually uploaded to our S3 bucket.

alpha_utils.py

Contains utility functions for grabbing data from Alpha Vantage API.