Module 1: Python Foundations
Set up your environment and master Python essentials for data work: data types, control flow, functions, modules, virtual environments, and notebooks. Work extensively with NumPy and Pandas for vectorized operations and tabular data handling.
Outcomes: Write clean, reusable Python, manipulate datasets efficiently, and build robust data pipelines.
Theory focus: We examine how Python manages memory and data structures under the hood (lists vs arrays, copying vs views), why vectorization eliminates Python-level loops, and the implications of immutability and pure functions on reproducibility. You will understand complexity trade‑offs of common transformations and the rationale behind using environments, dependency pinning, and notebooks as literate programming artifacts.
Module 2: Mathematics & Statistics
Probability, distributions, sampling, hypothesis testing, confidence intervals, correlation, regression, feature scaling, and regularization. Intuition-first with practical computation in Python.
Outcomes: Design sound experiments and choose statistically valid models.
Theory focus: We formalize statistical inference—sampling distributions, bias‑variance trade‑off, overfitting vs underfitting, and the geometry of least squares. You will study assumptions behind parametric tests, how violations manifest in diagnostics, and when to pivot to non‑parametric or resampling methods (bootstrap, permutation tests).
Module 3: SQL & Databases
Master SQL queries, joins, window functions, CTEs, and performance basics. Work with transactional (PostgreSQL/MySQL) and NoSQL stores. Build end-to-end data ingestion tasks.
Outcomes: Extract and prepare data reliably from real databases.
Theory focus: Learn relational algebra foundations powering SQL optimizers, normalization vs denormalization, indexing strategies (B‑trees, hash), and cardinality estimation. We also compare OLTP vs OLAP workloads, transaction isolation, and consistency models relevant to analytics engineering.
Module 4: Data Wrangling & EDA
Cleaning missing values, outlier handling, feature engineering, encoding, and pipelines. Perform exploratory analysis, profiling, and data quality checks.
Outcomes: Build explainable narratives from raw datasets.
Theory focus: Understand missingness mechanisms (MCAR, MAR, MNAR), robust estimators, leakage sources, and reproducible data contracts. We emphasize exploratory workflows as hypothesis generation, emphasizing visualization as a modeling tool rather than mere reporting.
Module 5: Data Visualization
Matplotlib, Seaborn, Plotly, and dashboards. Visual grammar, storytelling, and KPI design. Create interactive visualizations for stakeholders.
Outcomes: Communicate insights clearly and persuasively.
Theory focus: We cover perceptual principles (pre‑attentive attributes, color theory for data), Cleveland–McGill rankings, and the grammar of graphics. You will learn to align chart selection with data types and cognitive load to avoid misinterpretation and chartjunk.
Module 6: Machine Learning
Supervised and unsupervised learning with Scikit-learn: linear/logistic regression, trees, ensembles, clustering, model selection, cross-validation, and metrics. Handle imbalance and leakage.
Outcomes: Train, tune, and evaluate production-ready ML models.
Theory focus: We derive objective functions and regularizers, study capacity control, generalization bounds at a high level, and metric selection by problem formulation (ranking vs classification vs regression). We detail validation schemes under temporal and grouped dependencies.
Module 7: Deep Learning
Neural networks with TensorFlow/PyTorch, CNNs for vision, RNN/LSTM/Transformers for sequence, transfer learning, fine-tuning, and experiment tracking.
Outcomes: Build and optimize deep models for real use cases.
Theory focus: Understand gradient‑based optimization, initialization, vanishing/exploding gradients, normalization layers, and inductive biases of architectures (convolution, recurrence, attention). We examine regularization via dropout, weight decay, and data augmentation.
Module 8: NLP & Generative AI
Text cleaning, embeddings, topic modeling, sequence models, and LLM workflows. Use modern vector stores and prompt engineering for applied NLP.
Outcomes: Deliver chatbots, summarizers, and search systems.
Theory focus: Distributional semantics, subword tokenization, contextual embeddings, and retrieval‑augmented generation principles. We analyze evaluation pitfalls (hallucination, bias), safety guardrails, and prompt/system design patterns.
Module 9: Big Data & Cloud
Apache Spark for distributed computing, data pipelines, and streaming. Use AWS/GCP/Azure for storage, compute, and managed ML services.
Outcomes: Process large-scale datasets cost‑effectively.
Theory focus: We discuss distributed systems fundamentals—partitioning, shuffles, spill, and fault tolerance (lineage), plus cost models in cloud environments. Learn storage formats (Parquet/ORC), compression, and pruning for performance.
Module 10: MLOps & Deployment
Model packaging, APIs, CI/CD, monitoring, drift detection, and lifecycle management. Build dashboards and services for stakeholders.
Outcomes: Ship reliable ML systems from notebook to production.
Theory focus: We frame the model lifecycle as a socio‑technical system: data versioning, model lineage, observability (data/feature/concept drift), and feedback loops. Study deployment patterns (batch, real‑time, streaming) and reliability through SLIs/SLOs.
