Data Science and Machine Learning: Making Data-Driven Decisions

Weeks 1-2: Foundations of Data Science

In the first two weeks, we will cover the foundational concepts for data science that form the building blocks of the course and will help you sail through the rest of the journey with ease.

Python for Data Science

1 Case Study

1 Case Study

Python, for data scientists and machine learning specialists, is a lingua franca owing to the immense promise of this widely-used programming language. To strengthen your Python foundations, this module focuses on NumPy, Pandas and Data Visualization.

• Numpy

A Python package for scientific computing that enables one to work with multi-dimensional arrays and matrices.

• Pandas

An open-source and powerful library in Python that is used to analyse and manipulate data.

• Data Visualization

Dealing with the graphic representation of data, that is very effective in generating insights from data by using matplotlib, seaborn, etc. libraries.

Statistics for Data Science

1 Case Study

1 Case Study

This week will help you understand the role of statistics in helping organizations take effective decisions, learn its most widely-used tools and learn to solve business problems using analysis, data interpretation and experiments. It will cover the following topics:

• Descriptive Statistics

It gives you the basic measures of a statistical summary of the data.

• Inferential Statistics

This will explore the areas of distributions and parameter estimation, ultimately allowing you to make inferences from the data.

Week 3: Making Sense of Unstructured Data

In this week, you will learn about how to apply different ML techniques to discover patterns & insights in the unstructured data.

Introduction

Here, you will learn about one of the important aspects of ML - Unsupervised Learning.

• What is unsupervised learning, and why is it challenging?
• Examples of unsupervised learning

Clustering

2 Case Studies

2 Case Studies

Clustering is an unsupervised learning technique to group similar sets of data points. Here, you will learn about the widely used clustering techniques i.e. K-means clustering.

• What is clustering?
• When to use clustering?
• K-means preliminaries
• The K-means algorithm
• How to evaluate clustering?
• Beyond K-means: What really makes a cluster?
• Beyond K-means: Other notions of distance
• Beyond K-means: Data and pre-processing
• Beyond K-means: Big data and nonparametric Bayes
• Beyond clustering

Spectral Clustering, Components, and Embeddings

2 Case Studies

2 Case Studies

Spectral clustering is one of the most popular techniques when it comes to cluster graphs & networks. Here, you will learn about the spectral, modularity clustering, and PCA algorithm.

• What if we do not have features to describe the data or not all are meaningful?
• Finding the principal components in data and applications
• The magic of eigenvectors I
• Clustering in graphs and networks
• Features from graphs: The magic of eigenvectors II
• Spectral clustering
• Modularity Clustering
• Embeddings: New features and their meaning

Week 4: Regression and Prediction

In this week, you will explore the classical and modern regression methods for prediction and inferential purposes.

Classical Linear and Nonlinear Regression and Extensions

2 Case Studies

2 Case Studies

Here, you will learn about linear and nonlinear regression together with their extensions, including the important case of logistic regression for binary classification and causal inference where the goal is to understand the effects of actively manipulating a variable as opposed to passively measuring it.

• Linear regression with one and several variables
• Linear regression for prediction
• Linear regression for causal inference
• Logistic and other types of nonlinear regression

Modern Regression with High-Dimensional Data

1 Case Studies

2 Case Studies

Here, you will learn about the modern regression with high-dimensional data, or finding a needle in a haystack: for large datasets, it becomes necessary to sort out which variables are relevant for prediction and which are not. Recent years have witnessed the development of new statistical techniques, such as Lasso or Random Forests, that are computationally superior to large datasets and that automatically select relevant data.

• Making good predictions with high-dimensional data
• Avoiding overfitting by validation and cross-validation
• Regularization by Lasso, Ridge, and their modifications
• Regression Trees, Random Forest, Boosted Trees

The Use of Modern Regression for Causal Inference

2 Case Studies

2 Case Studies

This part will cover regression and causal inference to explain why “correlation does not imply causation” and how we can overcome this intrinsic limitation of regression by resorting to randomized control studies or controlling for confounding.

• Randomized Control Trials
• Observational Studies with Confounding

Week 5: Classification and Hypothesis Testing

In this week, you will learn about the basics of anomaly detection, classification, and fundamentals of hypothesis testing, which is the formalization of scientific inquiry. This delicate statistical setup obeys a certain set of rules that will be explained and put in context with classification.

Hypothesis Testing and Classification

1 Case Study

1 Case Study
• What are anomalies? What is fraud? Spams?
• Binary Classification: False Positive/Negative, Precision / Recall, F1-Score
• Logistic and Probit regression: Statistical binary classification
• Hypothesis testing: Ratio Test and Neyman-Pearson p-values: Confidence
• Support vector machine: Non-statistical classifier
• Perceptron: Simple classifier with elegant interpretation

Week 6: Deep Learning

Deep learning has been emerging as a driving force in the ongoing technological revolution. The essence of Deep Learning lies in its ability to imitate the human brain in processing data for various purposes, that too without any human supervision. Neural networks are at the heart of this technology. This week will take you beyond traditional ML and into the realm of Neural Networks and Deep Learning. You’ll learn how Deep Learning can be successfully applied to areas such as Computer Vision, and more.

Deep Learning

1 Case Study

1 Case Study
• What is image classification? Introduce ImageNet and show examples
• Classification using a single linear threshold (perceptron)
• Hierarchical representations
• Fitting parameters using back-propagation
• Non-convex functions
• How interpret-able are its features?
• Manipulating deep nets (ostrich example)
• Transfer learning
• Other applications I: Speech recognition
• Other applications II: Natural language processing

Week 7 - Learning Break

Week 8: Recommendation Systems

As organizations increasingly lean towards data-driven approaches, an understanding of recommendation systems can help not only data science experts but also professionals in other areas such as marketing who, too, are expected to be data literate today. Learn why recommendation systems are now everywhere and some insight on what is required to build a good recommendation system by covering statistical modeling and algorithms.

Recommendations and Ranking

1 Case Study

1 Case Study

Recommendation System algorithms, simply put, suggest relevant items to users - explaining the trends of their usage across a range of industries and their central role in revenue generation.

• What does a recommendation system do?
• So what is the recommendation prediction problem? And what data do we have?
• Using population averages
• Using population comparisons and ranking

Collaborative Filtering

1 Case Study

1 Case Study

Collaborative filtering is an aspect of recommendation systems with which we interact quite frequently. Upon collecting data on preferences of multiple users, collaborative filtering makes predictions for the choice of a particular user.

• Personalization using collaborative filtering using similar users
• Personalization using collaborative filtering using similar items
• Personalization using collaborative filtering using similar users and items

Personalized Recommendations

1 Case Study

1 Case Study

As suggested by the name itself, personalized recommendations work to filter out recommendations which are personally relevant for a user, based on their browsing trends, etc.

• Personalization using comparisons, rankings, and users-items
• Hidden Markov Model / Neural Nets, Bipartite graph, and graphical model
• Using side-information
• 20 questions and active learning
• Building a system: Algorithmic and system challenges

Week 9: Networking and Graphical Models

In this week, you will get a systematic overview of methods for analyzing large networks, determining important structures in such networks, and inferring missing data in networks. An emphasis is placed on graphical models, both, as a powerful way to model network processes and to facilitate efficient statistical computation.

Introduction

Here, you will get to know what networks are and how we can represent networks, with their practical use-cases around us.

• Introduction to networks
• Examples of networks
• Representation of networks

Networks

1 Case Study

1 Case Study

Here, you will learn about the common descriptive measures of a network, such as a centrality, closeness, & betweenness, and standard stochastic models for networks, such as Erdos-Renyi, preferential attachment, infection models, notions of influence, etc.

• Centrality measures: degree, eigenvector, and page-rank
• Closeness and betweenness centrality
• Degree distribution, clustering, and small world
• Network models: Erdos-Renyi, configuration model, preferential attachment
• Stochastic models on networks for spread of viruses or ideas
• Influence maximization

Graphical Models

1 Case Study

1 Case Study

Here, you will get to know how to use graphical models to estimate and display a network of interactions.

• Undirected graphical models
• Ising and Gaussian models
• Learning graphical models from data
• Directed graphical models
• V-structures, “explaining away,” and learning directed graphical models
• Inference in graphical models: Marginals and message passing
• Hidden Markov Model (HMM)
• Kalman filter

Week 10: Predictive Analytics

In this week, you will learn about some practical examples of temporal data sources and how we can begin to understand them. Then, you will dive into several strategies for feature extraction, including Deep Feature Synthesis with primitives and stacking. Finally, you will look toward models for the real world and how to ensure they successfully predict future data.

Predictive Modeling for Temporal Data

1 Case Study

1 Case Study

Here, you will learn about the structure of temporal data and how can we clearly define training inputs and outputs for prediction?

• Prediction Engineering

Feature Engineering

1 Case Study

1 Case Study

In this part, you will know how to utilize feature engineering techniques to extract meaningful insights from temporal data? What are effective strategies for evaluating model performance and preparing to deploy it in the real world?

• Introduction
• Feature Types
• Deep Feature Synthesis: Primitives and Algorithms
• Deep Feature Synthesis: Stacking

Certificate of Completion from MIT IDSS.

Upon successful completion of the program, you will receive one of the best data science professional certificates out there, for it will be from MIT Institute for Data, Systems, and Society (IDSS).