This blog focuses on using OpenAI Canvas, a tool that enables seamless collaboration and powerful AI-driven insights for data analysis. We’ll guide you through its features using a simple Diamonds dataset as an example. Follow these steps to explore how OpenAI Canvas can streamline your workflow.
Step 1: Access OpenAI Canvas
Visit chatgpt.com or chat.com.
Log in to your OpenAI account.
Free accounts have limited features.
Consider upgrading to a Plus account ($20/month) for extensive analysis and report generation.
Once logged in, locate the Canvas feature by clicking on the menu in the top navigation bar, and then selecting 'Canvas' from the dropdown options.
Step 2: Upload Your Dataset
Prepare your dataset. For this example, we’ll use a CSV file named diamonds.csv, which is available at https://github.com/lbsocial/data-analysis-with-generative-ai/blob/main/diamonds.csv.
The dataset includes details like diamond ID, weight, color, clarity, and price.
Upload the dataset by clicking the “Upload File” button in OpenAI Canvas.
Once uploaded, the dataset preview will appear in the left panel.
Step 3: Generate a Summary
Ask OpenAI Canvas to summarize the dataset:
Use a command like: “Write a short summary of the uploaded dataset.”
The AI will:
Analyze the dataset.
Provide an overview of columns, data types, and key insights.
Review the generated report, which appears on the right panel.
Step 4: Modify the Report
Select any section of the report for editing.
Ask the AI to:
Remove unnecessary parts.
Add new details (e.g., explanations of columns).
Customize formatting:
Change font styles.
Add headings or bullet points for clarity.
Step 5: Perform Advanced Analysis
Explore patterns in the data by asking specific questions:
Example: “What is the correlation between diamond weight and price?”
Use AI to:
Create statistical summaries.
Generate visualizations such as bar charts or scatter plots.
Identify anomalies or trends:
Ask the AI to explain unusual findings.
Step 6: Visualize Data
Request visual representations:
Example: “Create bar charts for categorical columns like color and clarity.”
The AI will:
Generate charts using Python libraries (e.g., Matplotlib).
Embed these visualizations in the report using markdown syntax.
Step 7: Enhance the Report
Add footnotes for additional context:
Example: “Search online for more information about the diamond grading agencies in the dataset.”
Refine the writing style:
Ask the AI to adjust the tone or reading level.
Insert images or charts directly into the report for a polished finish.
Step 8: Export the Report
Download the final report:
As a markdown file or HTML with embedded charts.
Ensure all visualizations and content are properly displayed in the exported version.
If necessary, retry the export to include missing elements.
Final Thoughts
OpenAI Canvas streamlines data analysis, combining AI capabilities with user-friendly tools. It simplifies the process of generating, refining, and presenting insights, saving valuable time. While occasional limitations exist, such as missing visualizations in exports, the benefits far outweigh the challenges.
Whether you are a beginner or a seasoned data analyst, OpenAI Canvas provides tools for collaborative and efficient data analysis. Remember, AI can make mistakes, so verify its outputs and avoid using sensitive information in your analysis. Share your results while acknowledging the assistance of AI to maintain transparency and encourage informed discussions.
References
Karina Nguyen. “Collaborative Writing and Coding with OpenAI Canvas.” DeepLearning.AI. Accessed January 5, 2025. https://www.deeplearning.ai/short-courses/collaborative-writing-and-coding-with-openai-canvas/
Comments