Review of Using ChatGPT for 200% Work Automation (feat. Python)

I automated repetitive data cleaning tasks at work using ChatGPT API + Python.

  • First, wrote a prompt explaining the desired output format to GPT-4
  • Read an Excel file with a Python script, then called the API
  • Wrote the results back to Excel

It took half a day to set up initially, but now I'm saving 4 hours every week. I once again realized the importance of prompt engineering.

I recommend this to anyone with similar concerns.

by 클라우드러버430

10 answers

Wow, that's amazing... I want to try this too.

by 문과출신개발자661 · ▲0

Crafting prompts is harder than I thought. Can you give me an example of how you structured it?

by 무한도전러56 · ▲0

Well, the key is how much the API cost will be. Even if it saves 4 hours, is it really worth it?

by 지나가던행인979 · ▲0

Agreed. Prompt engineering skills are really important.

by 주말개발자20 · ▲0

I also tried automating Excel cleanup, but there were more edge cases than I thought, so in the end, a human had to refine it.

by 궁금한사람483 · ▲0

Recommendation. If you really want to save time, the Python+GPT combo is the best.

by AI덕후194 · ▲0

Didn't someone already talk about that a while ago? But the review is helpful.

by 월급루팡341 · ▲0

Are there any security concerns? I'm a bit reluctant to send company data to an API.

by 초보개발자833 · ▲0

Oh, this is totally my situation. I should give it a try this weekend too.

by 주말개발자563 · ▲0

Actually, there are many times when doing it manually is more accurate than automation. Especially when the data is messy, GPT sometimes spouts nonsense.

by 알고리즘고수381 · ▲0