Events

Home >> Events

🍕Python & Pizza (Fall 2025)

Event 1

This hands-on session is all about exploring how we can use Python and Large Language Models (LLMs) to analyze online discussions about ChatGPT. You’ll get to see how data from Reddit can be collected, analyzed, and turned into meaningful visual insights comparing the perspectives of professors and students.

📂 Access the Workshop Materials

I have prepared a Google Drive folder with everything you need to follow along:

📓 Colab Notebook – ready-to-run code for scraping, analyzing through GPT, and visualizing Reddit posts.
🧾 Sample Datasets – posts from professors’ and students’ related subreddits (about ChatGPT).
📊 Output Examples – labeled CSV files and example plots.

👉 Click here to access the Google Drive folder

🔧 What You’ll Do

Use Python in Colab (no installation required).
Log into your gmail account
Copy the Google drive folder I shared above in your Google drive before running
- you can download, extract, and upload to your drive or
- Just copy – this will add a “Copy of” tag in the name of the file which you will need to remove
Label Reddit posts with sentiment, stance, and themes using GPT-4o-mini. (Already done for you)
Run the cells to create visualizations like:
- Bar charts of sentiment and stance
- Top themes chart for professors and students
- Word clouds of themes
- Theme co-occurrence networks etc.
When you go back home manually evaluate the posts and the labels to check how gpt-4o-mini performed on the data
Zoom Link (in case you want to remotely join): https://suny-ow-edu.zoom.us/j/7222550367

💬 Questions?

For more details about the Data Science program at SUNY Old Westbury, or to connect about future events, or bug reporting for the given codes, contact Dr. Shebuti Rayana at rayanas@oldwestbury.edu

✨ Grab some pizza, open the notebook, and start discovering insights from the online conversation about chatGPT using GPT ;)!

Event 2 – Thursday, October 30th

from 5:30 PM – 7:00 PM

📂 Access the Workshop Materials

📓 Colab Notebook – ready-to-run code loading, analyzing data, and creating plots.
🧾 Sample Datasets – multiple climate datasets from NASA, and Berkley.
📊 Output – climate plots.

👉 Click here to access the Google Drive

Folder

🔧 What You’ll Do

Use Python in Colab (no installation required).
Log into your gmail account
Copy the Google drive folder I shared above in your Google drive before running
- you can download, extract, and upload the flies to your drive or
- Just copy – this will add a “Copy of” tag in the name of the file which you will need to remove
- If you do not plan to run the cells to regenerate the charts you can only copy the notebook .ipynb file
Analysis and visualizations are already done for you
- The per-country temperature data in the temperature-data folder may be a little large – if you do not regenerate the globe with animation – you do not need to copy this folder to your drive
Your task is to complete building the data story that I initiated in my presentation, by
- Following the storytelling arc – add the falling action and resolution to the story – you can make 1-5 slides for that. Hint: find a dataset for sea level increase, etc. Or you can find some charts/graphs online to add to the story.
Zoom Link (in case you want to remotely join): https://suny-ow-edu.zoom.us/j/7222550367

Event 3 – Wednesday, December 3rd from 1:30 PM – 2:30 PM

🍕 Python, Pizza and Personality: Build Your Own Fun Chatbot

Hands-on Event
Instructor: Dr. Shebuti Rayana, Math/CIS, SUNY at Old Westbury

🎯 Goal

Create your own AI-powered personality chatbot using Hugging Face Spaces — no installation, no coding environment setup.

Step-by-Step Guide

🧩 Step 1. Sign in to Hugging Face

Go to https://huggingface.co
Click Sign Up or Log In (use Google or email).
Once logged in, click your profile icon → New Space.

🧠 Step 2. Create a new Space

Name your space: e.g., pizza-bot, or my-fun-chatbot.
License: Choose apache 2.0
Select SDK: Choose Gradio.
Gradio template: Leave blank (None) — do not choose “Chatbot”.
Space Hardware: Choose CPU basic
Visibility: Public (so you can share it).
Click Create Space.

🗂️ Step 3. Add your three files

You will find three files in this link: https://drive.google.com/drive/folders/1E6grJ5KdpG5kPeX-MnNz-OF5qL54V8wJ?usp=sharing

app.py — main Python code for your chatbot
requirements.txt — list of Python libraries
README.md — configuration info [Hugging Face will automatically create this file you can just edit the file and add info from the given file]

Upload them directly into your Space (root directory).
To do this: click the Files tab → +Contribute → Upload files.

⚙️ Step 4. Let it build

The Space will automatically start building the environment.
This may take 2–3 minutes (watch the logs if you’re curious).
Once complete, your chatbot will appear live!
Following is a screenshot of the chatbot

🎨 Step 5. Personalize your chatbot

Open app.py in the Hugging Face online editor.
Find the section labeled PERSONAS, for example:

“Professor Pizza”: “You are Professor Pizza, the witty mascot of the Python & Pizza events at SUNY Old Westbury. You speak in warm, funny pizza metaphors when explaining programming or AI. Keep answers under 3 sentences and always stay kind and school-appropriate. If unsure, say: ‘I’m not sure—ask Dr. Rayana.'”,

Now add your own personality as a new item under PERSONAS!

Save the file — the Space will automatically restart and rebuild with your new personality.

💬 Step 6. Test your chatbot

Type your question in the message box.
Switch between different personalities using the dropdown.
Try changing the sliders:
- Temperature → controls creativity
- Top-p → controls vocabulary variety

Example prompts:

🍕 Professor Pizza

“Professor Pizza, can you explain machine learning using pizza ingredients?”
“What’s the secret sauce for writing good Python code?”
“If debugging were a pizza topping, what would it be and why?”
“How would you describe AI ethics to a slice of pepperoni?”
“What kind of pizza represents neural networks?”

🕵️ Detective Data

“Detective Data, I think my dataset’s gone missing — what’s the first clue you’d look for?”
“There’s a bug hiding in my Python code. How would you interrogate the suspects?”
“How do you know when your model’s accuracy is lying to you?”
“What’s your theory about why neural networks forget things?”
“Describe data cleaning as if it were a detective story.”

✍️ Study Buddy Poet

“Write a poem about overfitting in machine learning.”
“Can you explain data cleaning in a short rhyme?”
“What’s your poetic advice for staying calm during finals?”
“Turn ‘gradient descent’ into a cute two-line rhyme.”
“Make a short poem about friendship between data and algorithms.”

📚 OW Librarian

“Librarian, what’s the official OW policy on missing semicolons?”
“Could you summarize AI ethics as stated in the ‘OW Student Handbook’?”
“How should a student approach debugging respectfully, according to OW guidelines?”
“What are the library rules for training a neural network overnight?”
“How does OW define good data hygiene?”

📢 Step 7. Share your chatbot

At the top of your Space page, look for the three-dot (⋮) menu next to the “Settings” button.
Click it and select “Embed This Space”.
A small window will pop up with your Direct link — copy that and share it.
Copy the link and share it with your friends to try it for fun!
You can also customize the Space with your own title and emoji in README.md.

💡 Bonus Challenges

Add a new slider for “Response Length”.
Add another personality with a different tone (e.g., sarcastic, poetic, motivational).
Change the model from Qwen2.5-0.5B-Instruct to another lightweight model on Hugging Face (like Falcon3-1B-Instruct or Phi-2).

🔧 Troubleshooting

If you see an error:

Check that you didn’t choose the Chatbot template during creation.
Make sure your main file is named app.py.
The top of README.md must start with the YAML block (the three --- lines).
Restart your Space manually (click ⋮ → Restart Space).

🎉 Done!

You’ve created your own personality chatbot hosted on Hugging Face — ready to share with the world!