Foundation of Data Science

Syllabus

1. Introduction to Data Science (3 hours)
   1. Overview of data science
   2. Jargons of data science
   3. Modern data ecosystem
   4. Data science lifecycle
   5. Trends, markets and applications of data science
   6. Tools and technologies in data science
   7. Data scientist and their roles
2. Mathematics for Data Science (10 hours)
   1. Introduction to linear algebra for data science
   2. Vectors, matrices and matrix factorization
   3. Gradient descent for optimization
   4. Introduction to probability and random variable
   5. Probability distributions: Normal, Bernoulli, Binomial, Poisson
   6. Descriptive and inferential statistics
   7. Central limit theorem and sample distribution concepts
   8. Normal approximation; hypothesis testing procedures: tests about the mean of a normal population
   9. The t-test, Z-tests for differences between two population means, the two-sample t-test, confidence interval for mean of normal population
   10. ANOVA
3. Data Understanding and Preprocessing (10 hours)
   1. Types of data: structured, unstructured, semi-structured
   2. Data preprocessing requirements
   3. Data sources and collection methods
   4. Data cleaning and preparation
   5. Data wrangling and associated tools
   6. Data enrichment, validation and publishing
   7. Data transformation and normalization
   8. Dimensionality reduction: linear factor model, principal component analysis (PCA)
4. Data Analysis (8 hours)
   1. Data analytics: descriptive, diagnostic, predictive and prescriptive analytics
   2. Exploratory data analysis using descriptive statistics
   3. Data visualization
   4. Data visualization techniques
   5. Principles of effective data visualization
   6. Feature engineering and other aspects of data manipulation
5. Regression and Predictive Modeling (5 hours)
   1. Empirical models, simple linear regression, MLE and least square estimator
   2. Multiple linear regression, matrix approach to multiple linear regression, polynomial regression models, categorical regressors, indicator variables, selection of variables and model building
   3. Logistic regression
6. Modeling and Validation Processes (6 hours)
   1. Introduction to machine learning
   2. Introduction to supervised, unsupervised and reinforcement learning
   3. Modeling process, training and validating model, cross validation methods, predicting new observations interpretation
   4. Measures for model performance and evaluation: classification accuracy, confusion matrix, sensitivity, specificity, precision, recall, F-score, ROC curve, clustering performance measures
7. Ethics and Recent Trends (3 hours)
   1. Ethical considerations in data science
   2. Data privacy regulations
   3. Responsible data usage
   4. The five Cs
   5. Future trends

Practicals

1. Get acquainted with data science tools and perform statistical analysis
2. Hypothesis tests (e.g., t-tests, Z-tests) on sample datasets to compare population means
3. Simulate and apply the central limit theorem (CLT) to demonstrate how sample distributions converge to a normal distribution
4. Perform data wrangling and ETL processes on a dataset, followed by exploratory data analysis (EDA)
5. Utilize tools to create effective data visualizations (line charts, bar charts, heat maps, box plots) to derive key insights
6. Implement feature extraction and selection techniques, including encoding methods like one-hot encoding and creating new features based on domain expertise
7. Develop a simple linear regression model, extend it to multiple linear regression, and visualize the regression line and residual plots
8. Apply logistic regression and evaluate the model using metrics such as accuracy, precision, recall, and the ROC curve
9. Apply K-means clustering and assess cluster quality using evaluation metrics like the silhouette score