Data Science for Managers offers a balance between theory and practice, with visualizations, demonstrations, exercises, case studies and projects.
Many managers have forgotten their advanced mathematics, so we emphasize visualizations of mathematical concepts instead of complicated proofs. Moreover, most participants are not professional programmers. We therefore present basic programming concepts and build up to complete solutions.
Novice programmers will learn how to read programming code provided in solutions; more advanced students will learn to build those solutions from the bottom-up using scikit-learn APIs. Teaching assistants are available to provide one-on-one assistance with practical problems. All participants will leave the course being able to build, evaluate, and work with real data and real models.
Data sampling, measurement, and wrangling
- Exploratory data analysis
- Data description, visualization, and graphing
- Bias, variance, and the bias-variance tradeoff
- Model validation and model cross-validation
- Hyperparameter tuning and information leakage
- Model evaluation and comparison
- Model weighting of costs and benefits
- Ensemble learning and meta-learning
- Predictive labeling and data augmentation
- Data-driven business models
- Big Data, Map-Reduce, and Spark
- Virtual machines and cloud computing
- Strategic planning for a digital transformation
- The management of talent and strategic human capital
Methods + Models
- Normalizing and standardizing data
- Linear and log-linear models
- Non-parametric models, splines and locally-linear models
- Nearest neighbor and similarity models
- Agglomerative clustering and K-means clustering
- Decision trees, bagging, boosting, and random forests
- Dimension reduction, PCA, t-SNE, and manifold projections
- Support vector machines
- Text as data and natural language processing (NLP)
- Word embeddings and latent topic modeling
- Feed-forward neural networks
- Convolutional neural networks
- Recurrent neural networks, LSTMs, bi-lateral LSTMs
- Generative adversarial networks
- Reinforcement learning
Preparation + Schedule
No prior training in Data Science is required to take Data Science for Managers, but some knowledge of Python will will help you to get the most out of the course.
At a minimum, we suggest completing the 7-hour Python tutorial by Kaggle before the course to ensure you get the most out of the examples. Doing so will help you follow along with the demos and project solutions covered in class.
- You should be familiar with linear algebra (although we use very little math)
- Be familiar with statistics (although we will review the basics)
- Be conversant in English (the course will be given in English)
- Bring a laptop (Mac, Windows, Linux, or Chromebook)
We will have several highly-qualified EPFL graduate students on-hand to help you throughout the course. Teaching assistants will be available to work with you one-on-one to answer questions about the programming code.
Data Science for Managers will be offered online, from 8:00 am to noon PST.
Learn more about the available data science courses taught by Professor Kenneth Younge
Data Science for Decision Makers
Specifically organized for time-crunched executives, this course summarizes the core concepts and methods of data science and then explores the strategy of digital transformation through specific examples, case studies, and group discussions.
Stay Updated About Data Science for Managers