Tepper · MBA

45-884 AI Methods for Social and Visual Data

Course I built

Using AI on text, networks, and images, and knowing when the output is ready for a decision.

I built this course for Tepper MBA students who need to use current AI methods on text, networks, images, and other unstructured data. We treat each model as a measuring instrument. Students should be able to say what it did, where it fails, and whether its output is good enough to act on. The labs are in Python, and we discuss ethics in the same labs as the code.

Topics

1. AI models and unstructured data: Python, JSON, calling a model's API, and prompt engineering
2. Natural language processing fundamentals
3. Text mining for business insight, with retrieval over company filings and vector databases
4. Computer vision fundamentals: object detection with YOLO, ResNet, and vision transformers
5. Mining visual data, with a case comparing Tesla Vision and Waymo's sensor fusion
6. Agentic AI and AI workflows, including building a first agent in n8n
7. Emerging tools: evaluations, LLM-as-judge, and MCP

How the course works

There are no prerequisites and no model training. Students use foundation models from OpenAI, Anthropic, Google, and Meta on text, images, and agent workflows, and the course teaches the Python, JSON, and API calls they need along the way.

Every lab compares models. Students classify product reviews with three models and weigh cost against accuracy, chain two agents that score semiconductor companies' risk from their filings, and in the computer vision class send one photograph through three generations of vision: YOLO boxes, ResNet labels, then a multimodal model that returns a shelf-restocking report as JSON.

Each class opens with AI Methods in the News, a student briefing on a product or a debate from the last six months. In Fall 2025 the agents module closed with Prasad Chalasani, co-founder of the Langroid agent framework.

Final projects

Final projects start from a real business question. Students show how their prompts changed and compare models before settling on one. In Summer 2026, seven of the nineteen projects were built with real companies or organizations, and three are already in use. Projects from Fall 2025 and Summer 2026, grouped by type, with one example each.

• Social listening and sentiment7
  Reddit sentiment through an aircraft maker's safety crisis, set against its stock price
• Finance and investing5
  An autonomous equity analyst triggered from a watchlist
• Career, hiring, and advising agents4
  A five-agent job-fit scorer with a skeptical hiring-manager critic
• Agentic workflow automation3
  Automating client outreach for a real-estate agent in n8n
• AI governance and ethics3
  Human oversight of AI in intelligence analysis
• Document and transcript extraction2
  Turning sales-call transcripts into buyer intelligence
• Computer vision in the field1
  An offline tool that helps bomb-disposal teams identify ordnance
• Mergers and corporate culture1
  Predicting merger success from culture fit in annual reports

How the course developed

I began building this course in December 2024 and first taught it in Fall 2025 to MBA and other master's students. It covers the data most business analytics courses skip: text, images, and the output of language models.

For Summer 2026 I rebuilt it for a part-time format, with one live session a week and a set of hands-on videos I recorded for each module. They run from a first Python lesson through retrieval over company filings, computer vision, and agent workflows. For Fall 2026 I moved agentic AI earlier in the term, and students now get a summary and a cleaned transcript after each class.

Offerings

• Fall 2025 full-time
• Fall 2025 online hybrid
• Summer 2026
• Fall 2026

Enrolled students get the notebooks and recordings on Canvas.