Adventures in python and algorithmic trading

Surepips

Junior member
These threads are exactly why I love amd have started joining these forums. I'm looking to broaden my knowledge on coding in general but specifically algorithmic coding, never used a bot before but have entertained the idea of programming my own for fun. Alot of great reference here. So thanks in advance
 

Sharky

Staff
Wow, it's been 3.5 years since I started this journal and 1.5 years since I last updated it. Well as you can probably guess, my progress both in terms of python and algorithmic trading got put on hold. Blame having a second child! My free time got super squeezed. Thankfully baby is now a 2 year old and recently started nursery, so I can finally commit my time again to updating the journal, but first I've got some long overdue housekeeping to do with T2W.

Stay tuned! 😅
 

Sharky

Staff
Housekeeping under way, and in the meantime, a new public commitment, to get a fully automated trading system running in python with real money up and running before the end of 2022. Much work to do, fun times ahead.. :cool:
 

Sharky

Staff
Python set up. Since we're going to be spending time in Mac's Terminal, we might as well upgrade to iTerm2. I found a good, up to date, guide here: https://towardsdatascience.com/the-ultimate-guide-to-your-terminal-makeover-e11f9b87ac99. I installed Oh-My-Zsh in order to install the PowerLevel10K theme with a little help from this tutorial: https://www.swtestacademy.com/customize-mac-terminal/ - note all of this is optional, I just like shiny things!

I also took the opportunity to upgrade Visual Studio Code, which is my text editor of choice having previously used Sublime Text.
 

Sharky

Staff
I used the free one for a long time, then finally ponied up for the paid version. I just kept coming across everyone in the blogosphere using VSC as their preferred editor, so I read up on some reasons people were switching eg. https://nickjanetakis.com/blog/switching-to-vscode-from-sublime-text and that convinced me to make the switch. To be quite honest most tools are more than adequate, inc ST - it's what you do with them that counts! If you're starting from scratch I'd say go straight to the most popular tools, as they've all got a learning curve.
 

Sharky

Staff
Warning to anyone attempting to set up a new M1 MacBook for algorithmic trading (or any sort of data science). You're going to need python packages like pandas and statsmodel, but these can't be installed on M1 Macs simply by using pip, the standard python package manager. After some digging, I found a workaround here: https://stackoverflow.com/questions/67602680/how-to-install-pandas-on-m1-mac

Bash:
python3 -m pip install virtualenv
virtualenv -p python3.8 venv
source venv/bin/activate
pip install --upgrade pip
pip install numpy cython
git clone --depth 1 https://github.com/pandas-dev/pandas.git
cd pandas
python3 setup.py install

This worked, you can do the same for statsmodel by cloning from https://github.com/statsmodels/statsmodels. In case you need a specific version you can always git checkout a particular version by finding the hash commit, see: https://tysonmaly.com/programming/git/how-to-clone-older-version-in-git/ and then just checkout the version you want before running the 'python3 setup.py install' command above.

Alternatively you can either use rosetta2 or miniforge/miniconda, as one data scientist ended up doing here: https://towardsdatascience.com/new-m1-who-dis-677e085baffd (and which I might end up doing myself).
 

Sharky

Staff
I've reached the conclusion that running any sort of python open source project with a bunch of package dependencies is most likely to fail on the M1 Mac simply because there aren't that many M1 configured versions available. Thus I think your best bet is to install the regular x86 version, and rely on Rosetta, it's slower but at least you should be able to run everything - which is fine if like me you're just running on your Mac for development purposes. I can always switch to using Docker or my linux laptop if the speed starts to become a problem.
 

Sharky

Staff
I've got all three of his books (that doesn't mean I've read them all cover to cover though!). Having said that, I definitely think Systematic Trading, or his newest book, Leveraged Trading, would be a better first book for most people. You can decide for yourself which one based on this handy comparison he wrote: https://qoppac.blogspot.com/2019/10/new-book-leveraged-trading.html (based on your account size and experience level).
 

Sharky

Staff
Following on from my last couple posts, I've made the decision to embrace Robert Carver's pysystemtrade as my algorithmic trading framework. Given his extraordinary generosity, not only do you have a free, open source platform but his complete futures trading system using both trend and non-trend following rules, all explained and dissected across his books and blog. No black box, fully customisable, and built from the ground up in Python. It ticks all my boxes.

Where to start then! So I decided to spend this week brushing up on a number of areas. It's amazing how quickly you forget things, but at least it's a lot quicker to pick up second (or third etc.) time round.

I wanted to accelerate my learning, and whilst in the past I've favoured learning from books (and continue to), the world has moved on, and these days I also really appreciate good quality educational video content, predominately on YouTube, to which I'm happy to stump up for the Premium subscription give how much time I spend on it (and my kids) during the week.

Whilst YouTube is great for finding quick videos on pretty much any topic, it lacks the structure of something like Udemy. I like how courses are broken down into sections and segments and it tracks my progress, even I choose to jump about different sections. I ended up signed up for a Udemy Pro Personal Plan Subscription that gives you a 7 day free trial (after which it's £26.99/month). This gives me access to tons of the best courses on the site and is a great way to throw myself into intensive learning. I'm covering a lot of different areas and I've narrowed down what I consider to be some of the best courses available on each subject on Udemy, based on ratings, what's included in the Pro plan and my own first impressions of each course. From these take your pick or better still combine them to really grasp the concepts learnt...

Basic Linux Administration

Version Control with Git and GitHub

Docker (optional, but handy!)

Python (lots to choose from, for now it's enough to get your head around the basic stuff)

Python Data Analysis (esp. numpy, pandas, jupyter)

Visual Studio Code (bonus extra - my preferred code editor!)
 

Sharky

Staff
Whilst not directly related to this topic, Andrea Neagoie also has this Javascript course which I think is maybe the first developer language that he learned.
It's also included as part of the Pro subscription. I like Neagole's teaching style and it looks like a great course syllabus if you want to get into web development. Python is great for data science, and can be used for web development with say Django and Flask, but node.js seems to be the way to go, server side, since you need to learn javascript anyway.
 

Sharky

Staff
Better books to read on systematic trading are Pardo and multiple books by Bandy.
I've got a copy of Design, Trading and Optimisation of Trading Systems by Pardo and hope to read it after I get through Leveraged Trading (first) and Systematic Trading both by Carver. For Bandy, any particular recommendation: https://www.amazon.co.uk/Books-Dr-Howard-B-Bandy/s?rh=n:266239,p_27:Dr+Howard+B+Bandy

Rob Carver, has quite a nice list of recommended reading here:
 
 
Top
AdBlock Detected

We get it, advertisements are annoying!

But it's thanks to our sponsors that access to Trade2Win remains free for all. By viewing our ads you help us pay our bills, so please support the site and disable your AdBlocker.

I've Disabled AdBlock    No Thanks