Requirements

• Python 3.x
• Required libraries listed in requirements.txt

Setup Instructions

1. Create a Virtual Environment: It is recommended to run the program in a virtual environment.

   python -m venv <directory>

2. Activate the Virtual Environment:

   • Windows:

     <directory>\Scripts\activate

   • macOS/Linux:

     source <directory>/bin/activate

3. Install the Required Libraries:

   pip install -r requirements.txt

4. Ensure the Whole Directory is Present: Make sure to run the file via the directory/open directory in terminal or IDE.

5. Enter API Key: Enter your API key from FinHub into FINHUB_API_KEY.txt.

6. Change Coin Symbol: Modify the coin symbol in scripts/data_writing.py line 51, e.g., 'BINANCE:BTCUSDT'.

7. Optional: Display Additional Stats: scripts/data_writing.py line 37 is optional, used to display additional stats (not used for ML tasks). Comment out when writing data for ML.

8. Change Frequency of Data Fetching: Adjust the frequency of data fetching in scripts/data_writing.py line 54.

Running the Program

When the data_writing.py starts running, processed_crypto_data.csv will start updating.

Visualization

1. Change Target Timezone: Change the target timezone in scripts/data_visualization.py line 13.

2. Run the Visualization Script: Run the file and open the port given in the terminal to view live visualization.

Jupyter Notebooks

There are Jupyter notebooks in the ML directory:

• Cleaning: Used just for cleaning data, not necessary for ML.
• XGB: Better at detecting dips in the data.
• LTSM: Better for detecting highs.

Run each notebook to view results.

Notes

• The models are not optimized for live data. '''