Open folder
Open the chat and set the working folder to your project root.
Workflows
simple machine learning projects
03
Prepare your capstone
Start with a small dataset and a clear goal. Use the agent to list files, read the data, and run a baseline script. Check the results in the chat cards. You can use "find datasets for data science projects" to practice your skills with a small dataset you own or may use under its licence.
List files
Run list_dir to see the current structure and verify your data exists.
Read data
Use read_file to inspect the first few lines of your CSV dataset.
Run baseline
Execute the baseline script in Shell and check the exit code for success.
Baseline script example
Run the baseline script and show the accuracy result.
- Shell
python baseline.pyexit 0 - read_file
baseline.py18 lines - write_file
baseline.pysaved - grep
baseline accuracy3 matches
waiting for your OK before the next stepRejectAccept
baseline.py
import csv
import sys
def main():
data = []
with open('data.csv') as f:
reader = csv.DictReader(f)
for row in reader:
data.append(row)
labels = [r['label'] for r in data]
most_common = max(set(labels), key=labels.count)
correct = sum(1 for r in data if r['label'] == most_common)
accuracy = correct / len(data)
print(f'Baseline accuracy: {accuracy:.2f}')
if __name__ == '__main__':
main()This pure Python script reads a CSV file, finds the most common label, and prints the baseline accuracy without using any third-party libraries.
Build your machine learning capstone project
Inspect data
List your project folder to confirm the dataset is present and use read_file to view the first few rows of your CSV data. Check the columns. This helps you understand the column names and data types before you start building. The agent shows the file content directly in the chat card for your review. You can use grep to find specific values or patterns in the data for simple machine learning projects. Verify the data looks clean and ready for training by checking for missing values. A first mini project for beginners often starts with inspecting these basic data structures. This step ensures you know exactly what you are working with for your machine learning pipeline.
Run training
Write a simple training script using your library and save it in the project folder, then use write_file to create the script and edit_file to fix any syntax errors. Run it. Run the script in the Shell tool to see the real output. Check the exit code to ensure the script finished successfully. Evaluate the model on your validation set and record the metrics. Save the results in a JSON log for reproducibility and clarity. Review the diff for every change when you use Ask first. Repeat this cycle for your machine learning capstone project while you choose a deep learning project.
