Optimize Crypto Trading Strategy (ML & Python)
Budget: $250 – $750 USD
Cryptocurrency Trading Strategy Optimization (Machine Learning + Python)
I'm working on a university-level individual project focused on developing and optimizing a cryptocurrency trading strategy using historical data and machine learning.
I will provide:
The initial code to download data from Binance
A Jupyter Notebook with the base machine learning model and trading logic
Your job is to support and help me implement and test different strategy parameters based on the following five key phases:
Project Tasks
1. Data Collection (20%)
Download 15-minute historical data from March 1st, 2024 onward using Binance API.
Work with two cryptocurrencies (which I will select) that differ in consensus mechanisms and blockchains.
You'll use the code I provide to collect and store this data.
2. Machine Learning Model (20%)
Apply the provided machine learning model to the collected historical data.
Use Expectancy as the performance metric to guide improvements.
Adjust the target variable and trading logic in the notebook.
3. Holding Period Optimization (20%)
Test several holding periods: 15min, 1h, 4h, 12h, 1d, 2d, 1w, 2w.
Modify the provided code to adapt .shift() logic and track how holding periods affect performance.
Identify which holding period generates the highest expectancy.
4. Take Profit Analysis (20%)
Label future return opportunities (e.g., top 15% return days).
Determine the optimal take-profit value that maximizes expectancy.
Adjust logic to label and identify strong future gains.
5. Stop Loss Strategy (20%)
Implement one or more custom stop-loss techniques, using academic or blog references (I'll share one to start with).
Test and compare performance of different stop-loss thresholds.
Modify the labeling in the notebook to reflect negative return days or stop-out conditions.
Deliverables
A complete Jupyter Notebook including all strategy optimization logic, charts, and final code.
A brief summary report (max 2 pages) comparing results across the two cryptocurrencies and summarizing which strategy worked best.
Skills Required
Strong Python (pandas, NumPy)
Experience using Binance API and handling financial time-series data
Understanding of basic machine learning and trading strategies
Bonus: knowledge of expectancy, backtesting, or trading metrics
I will provide you with:
Base data collection script (.py)
Base Jupyter notebook (.ipynb) with a working ML model
The chosen cryptocurrencies
I'm working on a university-level individual project focused on developing and optimizing a cryptocurrency trading strategy using historical data and machine learning.
I will provide:
The initial code to download data from Binance
A Jupyter Notebook with the base machine learning model and trading logic
Your job is to support and help me implement and test different strategy parameters based on the following five key phases:
Project Tasks
1. Data Collection (20%)
Download 15-minute historical data from March 1st, 2024 onward using Binance API.
Work with two cryptocurrencies (which I will select) that differ in consensus mechanisms and blockchains.
You'll use the code I provide to collect and store this data.
2. Machine Learning Model (20%)
Apply the provided machine learning model to the collected historical data.
Use Expectancy as the performance metric to guide improvements.
Adjust the target variable and trading logic in the notebook.
3. Holding Period Optimization (20%)
Test several holding periods: 15min, 1h, 4h, 12h, 1d, 2d, 1w, 2w.
Modify the provided code to adapt .shift() logic and track how holding periods affect performance.
Identify which holding period generates the highest expectancy.
4. Take Profit Analysis (20%)
Label future return opportunities (e.g., top 15% return days).
Determine the optimal take-profit value that maximizes expectancy.
Adjust logic to label and identify strong future gains.
5. Stop Loss Strategy (20%)
Implement one or more custom stop-loss techniques, using academic or blog references (I'll share one to start with).
Test and compare performance of different stop-loss thresholds.
Modify the labeling in the notebook to reflect negative return days or stop-out conditions.
Deliverables
A complete Jupyter Notebook including all strategy optimization logic, charts, and final code.
A brief summary report (max 2 pages) comparing results across the two cryptocurrencies and summarizing which strategy worked best.
Skills Required
Strong Python (pandas, NumPy)
Experience using Binance API and handling financial time-series data
Understanding of basic machine learning and trading strategies
Bonus: knowledge of expectancy, backtesting, or trading metrics
I will provide you with:
Base data collection script (.py)
Base Jupyter notebook (.ipynb) with a working ML model
The chosen cryptocurrencies