Proposal - Open Source Trading SW - Open AI Desk


Hello traders, programmers, analysts, e-coiners.

I would like to introduce you to an idea of mine: An generic, free & open source trading platform which enables you to automate you trades by an generic way.
You don't need to learn platform specific tasks and behavior for every provider you are making use of.

You might say: "One moment, there are still generic ways to trade". That's right, but please give me the chance to explain the benefits of my idea.

First of all, lets perform an inventory:

The traditional idea of money gets more and more replaced by virtual, decentralized new currencies: in general e-coins. To participate to these new currencies, you need a computer.

Many stock-exchanges in this area as well as traditional financial institutions providing us API's, which enables us to use third party software to execute trades or, at least, letting us query our account balance and market prices.

There are also software products, which enables us to perform trades and queries. These solutions are either local deployed programs or server based products behind an commercial supplier.

Why there is the need of a new solution:

The existing solutions have all a few disadvantages together: the locally deployed programs cannot be used on mobile devices and entail the necessary that the PC, which runs the program, is online and turned on all the time.

On the other hand, with the provider based solutions, as you can suggest, you have a vendor lock-in. Free open source currencies and vendor lock-In's are a bad match and doesn't comply with the spirit of open source.

How does a good solution should look like:

A good solution must be capable to be deployed locally as well as in the cloud. The handling must be comfortable with a desktop and mobile device. It must be scalable and easy to install. Last but not least, it must use of state of the art technology and it must be extensible.

The software must be free to use and open source.

My idea:

My architecture draft from an general point of view:

All parts of the software are Python based; there are many tutorials outside which picking up Python as the preferred language for analytics tasks.

Connectors: the connector layer enables to use different kinds of data sources. This part of the architecture is one which I can contribute to by myself. I wave written a C library for which I'm currently using as a base of a Python 3.x API (still under development).

Algorithmic Operations: This layer inherits all the useful libraries that makes Python to a powerful tool for analytics operations.

GUI: I could create a PyQT based GUI for this purpose, but this would lead into a one way street. To make this software more general, a web-based solution (like the I2P router console) would be a better solution. Unfortunately, I'm not into web development. I have chosen django because as far as I have heard you cant do much wrong with a django approach.

All parts are fit together in a docker container which allows an easy installation.

Thats my proposal: Feedback, suggestions, criticism welcome!