Portfolio balance

Streamlit application I use to rebalance my portfolio of securities (stocks and ETFs) each time I credit cash to my brokerage account. It finds the (quasi)-optimal set of operations to match a target allocation given current portfolio state (quantity, market prices) and of course, desired allocation, with the least amount of cash remaining (including my broker's fees).

Installation

Clone the repository by running the command:

git clone https://github.com/clabrugere/portfolio-balance.git

Run locally

In a new environment, install the required packages:

pip install -r requirements.txt

Launch the application:

streamlit run app.py

Run from a Docker image

Build the Docker image:

docker build -t portfolio-balance .

Instantiate the image in a new container:

docker run -d --name portfolio-balance-instance -p 80:80 portfolio-balance

You can now access the application in your browser: http://127.0.0.1

Note that you can change the fees_func in src/portfolio.py to adapt to your broker's fees logic.

Usage

Upload a csv file with the following columns:

• Asset: tickers (yahoo style "<stock>.<exchange>", e.g. "CW8.PA") of your current/desired positions
• Share: number of shares currently owned (0 for asset you would like to include in the portfolio)
• Weight: target allocation. The column must sum to 1

Note that the security needs to be available in yahoo finance.

Select the amount of cash you want to invest, and click "Balance"

How it works

It solves the non-linear problem with a non-linear constraint defined as:

where:

Because the independant variable is discrete, the objective and constraint non-linear (because $F$ is non linear), the optimal solution is searched using scipy implementation of Differential Evolution.

License

MIT