Enter your email to get the class slides

Discussion Summary

Getting Started with ChatGPT for Data Work

    • Using ChatGPT Plus
      • You need ChatGPT Plus (paid version) to access many features including Advanced Data Analysis
      • If you are operating on the older ChatGPT interface (shown in the recording), then you will need to turn on the ChatGPT Advanced Data Analysis feature.  (Following OpenAI’s Dev Day on November 6, 2023, they are rolling out a new ChatGPT interface that does not require you to turn on Advanced Data Analysis, it is automatically integrated.)
        • Click the three dots next to your username in the lower left
        • Go to Beta Features. Turn Advanced Data Analysis on
        • Click GPT-4 at the top (this will only be there if you are a paid subscriber). Select Advanced Data Analysis
      • Optional: Turn on plugins.  These are third party tools (still early in development / usefulness)
    • Tips For Effective Prompts
      • Short, specific prompts work better
      • Think of it as a conversation
      • Can save commonly used prompts to paste instead of retyping
    • Click the GPT-4 dropdown to select the model for the task at hand
      • Default – writing copy, summaries, viewing/analyzing images
      • Advanced Data Analysis – data work incl. summarizing, analyzing and cleaning data.  Writing code.
      • Bing – browsing and summarizing websites
      • DALL-E 3 – image generation
      • Plugins – mix of plugins from external companies.  Most in early phases.
    • Cleaning & Analyzing Data (with Advanced Analysis)
      • Upload an Excel file and have a conversation to get AI assistance summarizing and analyzing the data
      • Can help with data cleaning tasks like standardizing customer names
      • Useful for quick analysis when preparing for a meeting
      • Basic summary: Summarize by customer including total revenue and quantity.
      • More advanced: do cohort analysis to see customer trends over time
      • Always have the AI check its work and validate the results (will usually catch if it’s made a mistake)
      • You should also confirm totals make sense because it doesn’t always get everything right.
    • Creating Dummy Data (with Advanced Data Analysis)
      • Generate fake datasets for teaching, testing software, etc.
      • Ask for desired field names.  Iterate on the result with further prompts.  Export to Excel.
    • Summarizing Financials
      • Upload PDFs of models to summarize key data and assumptions
      • Can also upload images of financial statements for summary and analysis
      • Claude.ai does a great job at summarizing long documents and pdf financial statements (often better than ChatGPT)
    • Interview Prep
      • Get explanations of accounting concepts
      • Practice questions and sample answers.  Can use the audio feature on the phone to practice outloud.
    • Writing Code
      • Advanced Data Analysis mode can help write Python, SQL queries, etc.
      • Repl.it is an easy way to test the code
    • Creating Images
      • Use DALL-E in ChatGPT+ to generate images
    • Software Assistance
      • Get help using any software tool by asking ChatGPT

Key Limitations

  • Building models from scratch – can make mistakes
  • Converting PDF tables to Excel – hit or miss
  • Privacy issues with non-enterprise tools

Tips

  • Experiment with different models and prompts for each task
  • Short, specific prompts work best
  • Always have the AI check its work
  • Save commonly used prompts for reuse
  • Approach problems from multiple angles

Additional resources

To improve your Excel and Google Sheets skills, get better at double-checking with sumifs, countifs or pivot tables, join our Spreadsheet Fundamentals courses. 

You can also stay up-to-date on upcoming events as part of our AI and Data Analysis Learning Series by subscribing to our calendar.

Learn more about partnering with Recalc on AI training for your organization

Scroll to Top