---

👋 About this organization

This is the home of our class on Hugging Face. Students publish the datasets, models, and Spaces they build during the semester here, from the first EDA notebook to the final project demo.

The course gives you practical data science skills for business, economics, and entrepreneurship: framing problems, preparing data, building and evaluating models, and communicating insights that drive decisions. You'll work hands-on with industry-standard tools. No prior experience is required, and prior experience will only help.

🧭 The course in one line

Frame the problem → Get data → Explore → Model → Evaluate → Tell the story

📖 Full unit-by-unit syllabus

Unit	Topic	What we cover
1	Intro to DS & Lifecycle	Statistics vs. ML vs. Deep Learning · problem framing · DS roles and industry use cases · responsible DS: ethics, bias, reproducibility
2	Data, EDA & Visualization	Data sources (structured, unstructured, APIs, scraping) · cleaning: missing values, outliers, normalization · descriptive statistics and correlations · effective charts and dashboards
3	Classification	Supervised learning · logistic regression, decision trees, k-NN · accuracy, precision/recall, F1, ROC/AUC · fraud, spam, churn
4	Regression	Linear, Ridge and Lasso · residual diagnostics, over/underfitting · MSE, RMSE, MAE, R² · price prediction, forecasting
5	Clustering	Unsupervised learning · k-means, hierarchical, DBSCAN · choosing k (Elbow, Silhouette) · customer segmentation
6	Machine Learning (guest lecture)	Feature engineering and importance · train/test splits, cross-validation, data leakage · grid search · explainability
7	Recommendation Systems	RecSys algorithms · cold start · embeddings and similarity · search engines
8	Deep Learning	Neural networks vs. traditional ML · MLPs · embeddings · pre-trained models · tabular DL
9	Working with Text (NLP)	Tokens, vectors, embeddings · TF-IDF vs. transformers · prompting vs. fine-tuning · ROUGE · handling hallucinations
10	Working with Visual Data	How computers "see" · convolution · image classification, object detection, augmentation · pre-trained vision models
11	Multimodal Data (guest lecture)	Text + image + audio · cloning and deepfakes · synthetic data and augmentation
12	Course Summary	Recap · ethics: bias, privacy, responsible AI · AI agents, AutoML / AutoDS · retrospective
13–14	Presentations	Final project demos and oral defense

Some adjustments may occur. Guest-lecture dates are announced during the course, and advanced topics may adapt to the class's progress.

🎯 What you'll be able to do

• 🛠️ Build real AI applications and use pre-trained models on text, images, and more
• 📓 Produce clean, well-documented notebooks and reproducible pipelines
• 🔧 Engineer features, tune and evaluate models, and interpret results and metrics
• 📐 Apply core models: classification, regression, clustering, and beyond
• 🗣️ Communicate findings, visually and verbally, to stakeholders
• 🔁 Run an end-to-end workflow: framing → data → EDA → modeling → evaluation → storytelling

📦 What you'll find here

from datasets import load_dataset

# replace ORG_NAME with this organization's name
ds = load_dataset("ORG_NAME/dataset-name")

🤝 Publishing here: house rules

A good repo is one a stranger can understand and reproduce in five minutes.

1. Name it clearly, in lowercase with hyphens (for example eda-airbnb-listings or churn-classifier).
2. Write the card. Every dataset and model needs a README covering what it is, where the data came from, its license, how it was trained or evaluated, and its known limitations and biases.
3. Make it reproducible. Link the notebook or code, pin library versions, and fix random seeds.
4. Respect data and people. No private data or personal information, and no data you don't have the right to share. Credit your sources and keep their licenses.
5. Disclose your AI use (see the policy below). Add this block to your card:

## AI usage disclosure
- Tool(s) used: <e.g. Claude, ChatGPT, Gemini>
- What it was used for: <e.g. debugging, explaining a metric, drafting text>
- Example prompts: <the key prompts you used>

🤖 AI policy

Free use of generative AI is welcome, in line with the university's policy. When you use GAI tools in an assignment or exam, you must report it within your submission, stating how you used them: the type of tool, or the content of the prompts. Use that violates the university's guidelines may be treated as a disciplinary offense.

👥 Teaching team

Office hours: the lecturer is available before and after class, and online during the week on request. The tutor is available online during the week on request.

Course materials, lesson recordings, and submissions live on Moodle. Attendance follows the Entrepreneurship School's guidelines.

📚 Resources

🌐 Online tools & documentation

• Python.org
• scikit-learn
• pandas
• Kaggle: datasets and Learn
• Hugging Face: datasets, models, Spaces
• GeeksforGeeks: AI, DS, EDA, ML, DL, NLP, CV

✨ Generative AI tools

Groq · Claude · ChatGPT · Perplexity · Gemini

📘 Books

1. Cole Nussbaumer Knaflic (2015), Storytelling with Data
2. Kieran Healy (2018), Data Visualization: A Practical Introduction
3. Andriy Burkov (2019), The Hundred-Page Machine Learning Book
4. Ian Goodfellow, Yoshua Bengio, Aaron Courville (2015), Deep Learning
5. Tiffany Timbers, Trevor Campbell, Melissa Lee (2024), A First Introduction
6. Marek Gagolewski (2022), Minimalist Data Wrangling with Python
7. Jeffrey S. Saltz, Jeffrey M. Stanton (2024), An Introduction to Data Science
8. Shaun Ault, Soohyun Liao, Larry Musolino (2025), Principles of Data Science

🎓 Online courses

1. IBM Technology: AI Fundamentals
2. HarvardX: Data Science: Visualization
3. HarvardX: Data Science: Machine Learning
4. HarvardX: Machine Learning and AI with Python
5. HarvardX: Introduction to Data Science with Python
6. HarvardX: CS50's Introduction to Artificial Intelligence with Python
7. MIT (Fall 2016): Introduction to Computational Thinking and Data Science

---