ChatGPT for Advanced Data Analysis in Python notebook

Advanced Data Analysis (ADA) is a feature in ChatGPT that allows it to generate and execute Python code. This makes it possible to analyze data, create charts, process files (such as CSV or Excel), perform statistical calculations, and carry out various data-related tasks.

Key capabilities of Advanced Data Analysis:

• ✅ Generating Python Code – You can ask ChatGPT to write code for data analysis, visualization, or modeling.
• ✅ Executing Python Code – ChatGPT can run the code and return results, such as statistical analysis, numerical computations, or visualizations.
• ✅ Working with Files – You can upload files (CSV, Excel, JSON, etc.), and ChatGPT can read, process, and analyze them.
• ✅ Data Visualization – It can generate charts using Matplotlib, Seaborn, Plotly, and other libraries.
• ✅ Machine Learning – ChatGPT can help train and evaluate models, such as classifiers or regression models.
• ✅ Text Processing & NLP – It can analyze text, compute statistics, tokenize words, and more.

This feature makes ChatGPT function similarly to a Jupyter Notebook, but built directly into the chat interface. It allows you to quickly prototype analyses, test results, and experiment without setting up a local Python environment.

Example of Advanced Data Analysis in ChatGPT

Please click the plus icon in the chat box and upload CSV or Excel file with data. You can provide task to be performed in the prompt, in my example I asked ChatGPT to make a pair plot of features. You can see that there is Python code generated and executed. You can see a DataFrame displayed.

After several tries the ChatGPT constructed data visualization. Please click at the code box at the end of the response to check the code that was directly used to generate the plot.

The Python code is displayed in the box. It can't be edited by the user.

Limitations of Advanced Data Analysis (ADA) in ChatGPT

Although Advanced Data Analysis is powerful, it has several limitations:

1. Can't edit Python code manually

There is no option to manually tweak Python code generated by ChatGPT. Only code created by ChatGPT can be extecuted.

2. No Internet Access

ChatGPT cannot fetch real-time data from the web, access databases, online APIs, or scrape web pages. All data must be provided manually (e.g., uploaded files).

3. Limited Execution Time & Resources

The Python environment has limited execution time – complex computations may time out. Limited memory & processing power – it’s not optimized for running heavy machine learning models or deep learning training.

4. No Persistent Storage

ChatGPT does not retain files or execution history between sessions. If you upload a file, it is only available during the conversation. If the session is refreshed, all previous work is lost.

5. No GPU Support for ML/DL

While ChatGPT can help with machine learning, it does not have access to GPUs, so training deep learning models (e.g., TensorFlow/PyTorch) is impractical. Processing large datasets may be slow.

6. File Size Limitations

You can upload and analyze files, but very large files (hundreds of MBs or more) might not be processed due to memory constraints.

7. No Interactive UI Elements

Unlike Jupyter Notebook, ChatGPT does not support interactive widgets (e.g., sliders, buttons) and dynamic updates in plots (e.g., Plotly live charts).

8. Limited External Package Support

While many Python libraries work, some external dependencies (especially those requiring system-level installation) may not be available - user can't install new packages.

Advanced Data Analysis available in ChatGPT website is great for quick prototyping, small-scale data analysis, visualization, and simple machine learning tasks. It is not good choice for persistent projects, large-scale computations, deep learning training, or real-time web data access.

Using ChatGPT in Python notebook

ChatGPT can be used in Python notebook directly. There is a free desktop app called MLJAR Studio, that provides Python notebook environment and extensions to make data science tasks easier. The AI assistant in MLJAR Studio is powered by OpenAI ChatGPT model. The ChatGPT is provided in left panel and Python notebook on the right panel. The AI can produce Python code. The code is not automatically executed. User needs to review the code and accept its execution. The executed code is pasted into the Python notebook file. All executed code is saved in the file and can be later reused without AI assistance. Let's send a prompt to load a data.csv file:

We can ask AI assistant to create data visualizations:

AI assistant can ask user to install a new package in current Python environment. It won't install packages automatically.

The approach with accessing ChatGPT directly inside Python notebook has advantages over chat only interface.

1. Editable Python Code

🔹 In MLJAR Studio, the Python code generated by ChatGPT is not automatically executed.
🔹 The user can review, modify, and execute the code manually inside the notebook.
🚫 In the web-only interface, the generated code cannot be edited, and only the AI can execute it.

2. Persistent Code & Results

🔹 All executed code is saved in the notebook file and can be reused later.
🔹 This allows for iterative development without losing progress.
🚫 In the web-only interface, code and results are not saved—once the session is refreshed, all work is lost.

3. Full Python Environment Control

🔹 Users in MLJAR Studio can install new packages if needed.
🔹 The AI assistant suggests missing dependencies and allows manual installation.
🚫 The web interface has limited package support, and users cannot install new packages.

4. Better File Handling

🔹 In MLJAR Studio, files are loaded directly into the Python environment, making it easier to work with large datasets.
🔹 Users can process, transform, and store data persistently.
🚫 In the web interface, files are only available during the chat session and are lost after a refresh.

5. No Execution Time Limits

🔹 In MLJAR Studio, users control execution time and can run long computations.
🔹 There are no strict session limits like in the web version.
🚫 In the web interface, long-running computations may time out.

6. No Internet Access

🔹 In MLJAR Studio, Python execution has access to internet.
🔹 User can access real-time data from the web, access databases, scrape web and online APIs 🚫 In the web interface, there is no internet connection.

7. Ideal for Long-Term Projects

🔹 Since all executed code is saved in a notebook, users can continue analysis later without AI assistance.
🔹 This makes it ideal for machine learning pipelines and reproducible experiments.
🚫 In the web interface, once the chat session ends, all progress is lost.

Video showcasing MLJAR Studio used for data analysis:

Summary

ChatGPT has a feature called Advanced Data Analysis (ADA) that allows it to generate and run Python code. This helps users analyze data, create charts, process files, and perform machine learning tasks. The web version of ChatGPT provides a simple way to do these tasks, but it has several limitations.

One major issue with the web version is that users cannot edit the Python code before execution. ChatGPT runs the code automatically, and there is no way to modify or improve it manually. Another limitation is that files and results are not saved—once the session ends, everything is lost. The web version also does not allow installing new Python packages, which can be a problem for more advanced data analysis.

MLJAR Studio solves these problems by offering ChatGPT inside a Python notebook. Here, the AI suggests code, but users can review, edit, and run it manually. The executed code is saved in the notebook, making it easy to continue work later. MLJAR Studio also allows installing new Python packages, so users have full control over their environment.

Another advantage of MLJAR Studio is better handling of large files and long-running computations. The web version has strict time limits, while MLJAR Studio allows users to run more complex analyses without interruptions. It also provides a more flexible and interactive way to visualize data, as users can tweak and customize charts before execution.

Overall, MLJAR Studio is a better choice for serious data analysis and machine learning projects. It gives users more control, saves work for later use, supports larger datasets, and allows package installation. If you need to use ChatGPT for Python coding, MLJAR Studio offers a much more powerful and flexible experience than the web-only version.