DISCLAIMERS: I am a newbie, and this will sound like a rant. Because it is. I have done a lot of googling and read a few books on how to get into algo trading, and my experience has been frustrating. I am at the point where I am suffering from information overload and I feel the need to write down what I've accomplished so far (which isn't a lot). There's a lot of products out there. Lately I have been looking for data sources and there are tons of companies offering data sets and feeds. Some of them are very expensive. A few of them are cheap or free. Choosing one is not easy. And there's a different product offered for every stage of the development process. Need something for backtesting? Google gives you a million different products. Charting? Ditto. Brokers? Same thing. Education? All in one solutions? You get the idea. There's too many products to choose from. What happened to an "industry standard"? Errors! I'm a novice coder and most of the time the s**t just doesn't work! Sometimes this is my fault, yes. But there have been a few times where I wished that there was a way I could research the solution to the problem! Where is the good api documentation? Is the documentation updated and accurate? Where is the vibrant, active and helpful community?! I realize some products have great active communities, but which ones? And how do I find them? Some of you guys who are selling stuff in this sector need to make nicer looking apps because half of this stuff looks like it was made in 1995. I want to make stuff in python, please. what's with all the c# apis? That's all I can think of for now. Please school me and tell me I don't know what I'm talking about.
Trading is not just about simply pressing the BUY or SELL button. It is like trying to carve a worthless stone into a priceless gem. It takes years/decades (or infinity ), not weeks/months to be a trader.
From what I know, C# is faster than Python. So C# for execution and Python for backtesting and the like. As a noob myself, I feel your pain. Good luck.
That's a good analogy, comparison. The people who think trading is easy, and you use basic indicators of when to BUY or SELL....are the types of guys who wear a Rolex, or fake Rolex. Cheap, cliche, and laughable. While the truly wise, and in depth, and successful.....are like rare gemstone cutters.
It looks like you are trying to get on board too many subjects at once. You can pass the jungle of available information by making one step at a time. First plan your journey, what are you trying to develop? Could you scribble a few lines on what are your high-level goals? If you were treating it as a project those would be your milestones or epics.
Learn to trade first -- become consistently profitable, then look for ways to automate your process. If you want to algo trade, find a strategy that works first. The stuff that I use doesn't look fancy. My charts are built on Java Swing which I think came out around 1995. I use it because it makes money. I'll put my time into things that will make me more money / improve my efficiency executing trades. If I want to see something that looks good, I prefer to spend my time not looking at a computer. Python runs on an interpreter. The language is much slower than Java, C++, or C#. The only reason it is used for ML / HPC is because it can work with compiled highly-optimized libraries that are much more efficient (usually written in C or CUDA). Great for scripting and stuff commonly done in a Jupyter Notebook, not so great when you have a project that has several hundred or millions of lines of code.
Python can be sped up and even as standard it works for most, including me. Whether placing a trade takes 100ms or 250ms makes very little difference to me.
If you're a novice trader and a novice coder, you need to define a reason why you want to go algo? Algo IS pretty much the only way but i doubt your strategy will reflect that. So, ask yourself, why do you need algo trading in the first place? What edge do you have in front of others? Who's money are you going to take? Oh and if products are shitty, then you have all the opportunity to make a better one and get rich doing it C# is the top of the food chain. If everything else becomes so god damn efficient (and it will, even if it takes 100 years.) then speed is where the last edge lies.