StockNetworks

Exploring the use of graph networks through Neo4j for visualising relationships between stocks in the London Stock Exchange.

Structure

The project is built using Docker Containers through a Docker Compose.

1. App Container

Includes the functions and classes to:

• Read the data from the CloudIEX API
• Write it to the Postgres database
• Format it to obtain percentage change
• Calculate correlations
• Push back into the Neo4j database

2. Postgres Database

3. Neo4j Database

Execution

Requirements
Make sure there exists a file called vars.env with the correct credential information.

To start up the containers run:

docker-compose up -d

Then go into the 'app' container by using:

docker-compose exec app bash

Inside this container you can now execute the different commands using the CLI:

root@d0df06670440:/app# python3 main.py -h
usage: main.py [-h] {gs,p,cor,neo} ...

Stock exploration using Graph Networks

positional arguments:
  {gs,p,cor,neo}  Action type to perform
    gs            'gs' for getting Symbols from the API
    p             'p' for fetching prices from the API
    cor           + 'cor' for calculating the pairwise correlations between
                  the stocks, this command has 3 arguments: + '-sd' : start
                  date for price data (YYYY-MM-DD) + '-ed' : end data for
                  price data + '-ns' : number of stocks to correlate (these
                  are ordered by transaction volume over the previous period)
    neo           Add stock data to Neo4j database

optional arguments:
  -h, --help      show this help message and exit

The arguments for each of the functions can be found by adding the -h flag to the command. For example, to see arguments for the 'cor' function, use:

python3 main.py cor -h

Which returns:

root@d0df06670440:/app# python3 main.py cor -h
usage: main.py cor [-h] [-sd START_DATE] [-ed END_DATE] [-ns NUM_STOCKS]

optional arguments:
 -h, --help            show this help message and exit
 -sd START_DATE, --start-date START_DATE
                       Format is YYYY-MM-DD
 -ed END_DATE, --end-date END_DATE
                       Format is YYYY-MM-DD
 -ns NUM_STOCKS, --num-stocks NUM_STOCKS

To explore the data in the Postgres database use a Database IDE to connect to the Postgres database, with the following parameters:

• host: localhost
• port: 5432
  User, password and database should match those defined in the Docker-Compose file.

For access to the Neo4j database, use the built in GUI. It can be accessed from localhost:7474