Skip to content

Kantaphonkit/AutoGPT

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

🧪 AutoGPT vs Manual Data Analysis — Titanic Dataset

This project contains two experiments designed to explore and compare the capabilities of an AI Agent (AutoGPT) with a manually guided data analysis workflow. The Titanic dataset is used as the case study.

🚀 Experiment 1: AutoGPT Completes a Data Analysis Report

Objective: Experience the autonomous task planning and execution capabilities of an AI Agent.

Description: AutoGPT is given the goal:

"Please analyze this dataset and generate a report containing key statistical information and visualizations."

AutoGPT then automatically:

  • Loads the Titanic dataset

  • Calculates descriptive statistics

  • Plots relationships (e.g., survival rate vs gender)

  • Summarizes findings into a report

This experiment highlights the strengths and weaknesses of autonomous data analysis performed by an AI Agent.

🔍 Experiment 2: Manual vs AI-Guided Analysis Comparison

Objective: Develop an understanding of the capabilities and limitations of AI Agents by comparing them with a traditional manual analysis workflow.

Description: You manually perform the same analysis tasks and compare the results with AutoGPT:

✔️ Steps taken

✔️ Time consumed

✔️ Depth of insights

✔️ Accuracy and correctness

✔️ Efficiency and clarity

This comparison helps reveal where AI Agents can genuinely improve productivity and where human judgment is still needed.

📌 Summary

This project demonstrates:

How AutoGPT behaves as an autonomous data analyst

How its generated report compares to a human-made report

The practical strengths and current challenges of AI Agents in data analysis

Useful for learning about AI Agents, autonomous workflows, and human–AI collaboration in data science.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published