Sign inGet started
← Back to all guides

How to use ChatGPT for data science

By Nick Barth

Updated on March 6, 2024

Data science is a field that sits at the crossroads of statistics, computer science, and domain knowledge, using various tools to analyze and interpret complex data. Chatbot interfaces like OpenAI's ChatGPT offer an intuitive way to engage with data science tasks, from analysis to text generation. Here, we'll guide you through the main ways ChatGPT can be harnessed for data science applications, especially for users of platforms like Deepnote.

Using ChatGPT for data analysis

Exploratory Data Analysis (EDA)

  1. Understanding data distributions: Ask ChatGPT to explain the significance of data distributions and statistical measures like mean, median, mode, variance, and standard deviation.
  2. Hypothesis testing guidance: Use ChatGPT to get step-by-step instructions on how to perform hypothesis testing or to understand which statistical test fits your data.
  3. Data cleaning tips: Get advice on best practices for data preprocessing and cleaning, such as dealing with missing values or outliers.
  4. Feature selection strategies: Discuss with ChatGPT the principles of feature selection and how it impacts model performance.

Visual Data Insights

  1. Plot suggestions: Ask ChatGPT for recommendations on the type of visualizations that would best represent your data.
  2. Interpreting graphs: Describe your plots and visualizations to ChatGPT for insights on what they might indicate regarding your dataset.

Question answering

  1. Specific queries: ChatGPT can assist in answering specific questions about your dataset, such as "What does a high kurtosis value indicate about my data?"
  2. Modeling advice: For queries related to selecting and tuning data science models, ChatGPT can provide recommendations and explanations.

Text generation and Natural Language Processing (NLP) with ChatGPT

ChatGPT's robust NLP capabilities allow for the generation of text that can aid in data science workflows.

  1. Generating reports: Use ChatGPT to draft initial reports on data analysis, creating summaries of findings that can later be detailed or revised.
  2. Creating documentation: Get help writing documentation for your data science projects, ensuring clarity in methods and results.
  3. Data labeling: In supervised learning tasks, ChatGPT can assist in generating initial labels for textual data, which can then be fine-tuned by the data scientist.

Integration with Deepnote

Deepnote is a collaborative data science notebook platform that integrates seamlessly with Python code and Jupyter notebooks.

  1. Generating code: Ask ChatGPT to generate Python code snippets that you can run directly in Deepnote, whether for data manipulation, visualization, or model building.
  2. Debugging: If you hit a roadblock with your Deepnote notebook, describe the error to ChatGPT for troubleshooting tips and solutions.
  3. Workflow streamlining: Discuss with ChatGPT how to create more efficient data science pipelines and workflows within Deepnote.
  4. Deepnote collaboration: ChatGPT can explain the collaborative features of Deepnote and suggest ways to leverage them for team-based data science projects.

Final thoughts

For data scientists, Deepnote users, and AI enthusiasts, ChatGPT opens up exciting possibilities to enhance data science tasks. Whether it's through guiding exploratory data analysis, generating descriptive statistical text, or even aiding directly within the Deepnote ecosystem, tools like ChatGPT are reshaping how we approach data.

Remember that while ChatGPT is a powerful tool, it still requires human intuition and expertise to guide it effectively. Always validate the output and use your domain knowledge to interpret the results accurately.

Happy analyzing, and stay data-curious!


Note: All interactions with ChatGPT should comply with its usage policies, and the generated content should be fact-checked, especially when used in professional and data-critical environments.

Nick Barth

Product Engineer

Nick has been interested in data science ever since he recorded all his poops in spreadsheet, and found that on average, he pooped 1.41 times per day. When he isn't coding, or writing content, he spends his time enjoying various leisurely pursuits.

Follow Nick on LinkedIn and GitHub

That’s it, time to try Deepnote

Get started – it’s free
Book a demo



  • Integrations
  • Pricing
  • Documentation
  • Changelog
  • Security




  • Privacy
  • Terms

© Deepnote