Is this what a dev does?
In this series of articles, I'm going to take you through a little journey that I've decided to label as 'NotADev'.
Why, you may ask? Well, I've dabbled in Python scripts for some years whilst doing web2 ethical hacking and recently transitioning into web3 security research. Many moons ago, I was an analyst, building propensity models, forecasting, and performing various methods of customer analysis. So, I still have an interest in data, algorithms, and statistics.
Lately, I've tried my hand at trading, both in crypto and stocks & shares—only casually—but I have my hand in where I can on the stock market.
Over the past few months, I've caught the bug to do some dev work, starting off by building things myself and utilising AI for new ideas or code improvements. Then, Replit came out with Replit Agent, and I jumped at the opportunity to build some apps based on ideas I'd had, and I was truly amazed! Fast forward to today, I'm still using it, but I came across the new 'ChatGPT o1 preview' model and thought I'd test it to improve my already AI-written analysis bot for the stocks and shares market..
The idea.
I wanted to come up with a way of trading on the stock market in an automated fashion—saving me time analysing, monitoring prices, and buying or selling, all the usual stuff. So, I aimed to build a bot that would do this and then relay the results back to me via message, hoping to get the bot to trade for me too eventually.
Using free resources for data, ChatGPT (paid) for code generation, Telegram API for sending results, and some personal preferences, I began the quest to build a bot to run in the cloud
Requirements.
Automated extraction of historical stock market data
Incorporate trading indicators for signals
Utilise Machine Learning to predict future prices for weekly and daily intervals.
Aim for accuracy over 80% (personal preference).
Weekly analysis being the focus, with daily used as a run rate to capture volatility and adjust the weekly if necessary.
Build a back-testing model to test the algorithm/models historically
Save a summary of the results by date for each ticker
Message the signals on a daily basis
NTH (Nice To Have):
- Bot trading from signals on stock & shares market
My Setup
A primitive setup I’m sure, but since I’m not an actual developer, It’s suitable for the circumstances.
| What | Links | Comments |
| VS Code IDE | Visual Studio Code | Used for all my “Dev” and web3 security research. |
| ChatGPT |
- o1-preview
- o1-mini
- 4o | https://chatgpt.com/ | This is the AI I'll be using to generate the code, build the models, and algorithm. |
| Python 3 | Python | Language used for coding. I'm not advanced in this, and I'll lean on the AI for 95% of the code as I'm avoiding building it manually on purpose—also, it's a lot quicker than me. |
| Python Virtual Environment | - Virtual Environments | |
| How to Install | I always use a virtual environment for my Python coding. I encourage you to do the same—much cleaner, lfewer bugs and conflicts. | |
| Linode | https://www.linode.com | I use this service for cloud instances at an inexpensive cost.
This is my referral link; you'll receive a $100 60-day credit once you've added a valid payment method to your account |
| Wikipedia | List of S&P 500 companies - Wikipedia | List of the S&P 500 companies. I chose specific industries that I would trade in. |
| WSL2 | Install WSL2 | This is my go-to for my setup. I have a Windows laptop, but I use Linux 80% of the time. Linking this to my VS Code IDE is a must—can't work any other way |
So, that makes a start, see you in the next one.
pxng0lin.
![Article = [i for i in NotADev]](/_next/image?url=https%3A%2F%2Fcdn.hashnode.com%2Fres%2Fhashnode%2Fimage%2Fupload%2Fv1727480461686%2Fcbfa04ea-b838-40d3-8bc7-9e25a546a092.png&w=3840&q=75)
