Thread by @aakashgupta

🚨 BREAKING: Code Interpreter is FINALLY rolling out to all ChatGPT Plus users.

It's the most powerful feature OpenAI has released since GPT-4. It makes everyone a data analyst.

Here are 15 mind-blowing use cases of Code Interpreter:

1. Segment your customers in seconds

It takes a spreadsheet and then comes up with different segments of music markets - on its own.

This used to take hours of a human coding in R or Matlab.

2. Decompose seasonality simply via text

On its own, it figured out the exact seasonality in the price of bitcoin - in the blink of an eye.

This used to be an analysis you had to conduct.

Source: @TechMemeKing

3. Linear regression by itself

Simply asking Code Interpreter to come up with "interesting hypotheses" results in a well-constructed automated linear regression.

This used to be something you code via R or SAS after coming up with the hypothesis yourself.

Source: @emollick

4. Easy Geo Charts

You can just upload location data and get a GIF of the thing visualized - in this case, lighthouses in the US twinkling.

You used to have to connect this to expensive software like Tableau.

Source: @emollick

5. Basic descriptive charts in seconds

Just asking for "some basic visualizations" can get you all your data exploration steps that used to take hours in seconds.

You used to have to come up with the ideas yourself, even if the charts were easy to generate.

Source: @OpenAI

6. Graph public data without input

It can fetch data from public databases like the IMF and visualize it for you without any work.

This used to be a process of finding data, loading into your software, then formatting the chart. All is now done for you.

Source: @AiBreakfast

7. Automatic Radar Charts

It generated this hard-to-create chart by itself after analyzing this user's 300 hr Spotify playlist.

These used to only be available in certain software, and cumbersome to configure.

Source: @SHL0MS

8. Heatmaps With Ease

A CSV of SF crime data resulted in this heatmap with no guidance.

Source: @backus

9. Output the log chart automatically

Just ask ChatGPT to analyze a dataset and it'll figure out when a log transformation applies, and output it itself.

You used to have to do the transformation yourself and then chart it.

Source: @TechMemeKing

10. Cohort chart with no effort

Just upload data and it will build your cohorts and chart it in seconds.

This used to be several steps of grouping and charting. Now it's done automatically.

Source: @danshipper

11. Choose your clustering algorithm and debug

At scale clustering into groups of 100 used to be a tedious process in Python.

Code Interpreter does it all for you in seconds. And debugs its mistakes.

Source: @jmilinovich

12. Natural language querying to reduce stakeholder requests

Data scientists used to always get distracted with simple questions from stakeholders like "what's the average list price?"

Now they can just do it in Code Interpreter.

Source: @JimSpiewak

13. Plot mathematical functions

You used to have to go into a math tool to create a simple plot, and then define all the elements of the formula.

Code Interpreter will make assumptions and do it without any guidance.

Source: @EasyGuideAI

14. Clean data

Data cleaning used to be a tedious task filled with human decisions.

Code Interpreter will make intelligent decisions for you, like remove unnecessary columns.

Source: @olliethedev

15. Do all of the above in one place in seconds

In the future, everyone will be a data analyst.

Thanks to Code Interpreter.

Source: @Saboo_Shubham_

If you want more tips to help you use Code Interpreter, subscribe to the newsletter for my upcoming How To.

news.aakashg.com

I hope you found that helpful!

If you did,
1. Follow me @aakashg0 for tech breakdowns
2. Retweet the first tweet for others to see:

https://twitter.com/aakashg0/status/1677129060055674880