One session dedicated to all the terminology and evolution (example: PC to server to VM to VM in cloud). We will split this session to cover server, network, database, analytics, ML and AI.

What do we mean by data-driven decisioning? What has industry research established about businesses that are data-driven in their decision making vs. those that are not? Understand the difference between business intelligence (hindsight executive dashboard/reporting analytics), data science (hypothesis/diagnostics driven analytics), machine learning (predictive & prescriptive analytics) and artificial intelligence (cognitive intelligence).

What are the typical technologies and tools that data scientists use? Understand the difference between on-premise vs. cloud technology alternatives. We will learn how to setup a typical data science environment: both on-premise (on our laptops) as well as in the cloud employing a minimalist set of tools. Understand the difference between No-Code/GUI driven tools vs. Code-first approach.

Understand the evolution of Data Science-ML/AI methodologies from CRISP-DM to TDSP to CPMAI. Establish how these advanced analytics methodologies relate to Agile/Scrum (SAFe 4.0). We will demonstrate the importance of clearly understanding the business need before embarking on an advanced analytics exercise within the context of business process (batch vs. real-time), data assets and availability (structured, unstructured vs. streaming), the technology/application integration and success measures/KPIs.

Introduction to basic mathematical, statistical and computer science concepts that are frequently invoked by data scientists/machine learning engineers. Examples include: vectors, matrices, iterations, step-size, tolerance, statistical significance, inner and outer joins.

Understand what is involved with data collection and ingestion – from data profiling and manipulation to cleansing and validation, across numerical, categorical and boolean data types. We will also learn the importance of setting up train vs. validation vs. test samples – and typical methods employed to setup train vs. test (e.g. in-time vs. out-of-time samples).

Understand what is involved with data exploration – from target/outcome definition and hypothesis design to data visualization and setting the stage for feature engineering.

Understand what is involved with feature engineering – from data transformations and feature proxies to interpreting the relationship to the target/outcome and impact on production data preparation pipelines.

Understand the difference between supervised and unsupervised learning algorithms – the pros and cons of each, and guidelines for where to employ each category of algorithms.

Understand how to evaluate models, what are some of the most frequently used metrics employed in validating a model – the pros/cons of each metric, how to interpret the metrics – and how do these vary by the type of advanced analytic being pursued (classification vs. estimation vs. time-series vs. unstructured data/text/image analytics). Examine the difference between Model Validation and business Benefits Quantification (lift, incremental value, etc.).

Understand the relevance of model monitoring, governance and ethical ML/AI in the context of operationalizing advanced analytics models to production. What are some of the regulatory considerations that need to be considered (FCRA, GDPR/CCPA, HIPAA, etc.) and what types of model performance monitoring should be considered (data drift, model drift, decision performance, etc.).

Understand the difference between model deployment (batch, real-time vs. A/B testing) and model serving (associated technology architecture: server vs. serverless). Revisit CRISP-DM, TDSP/CPMAI to establish how DevOps and CI/CD fits into the analytics lifecycle and the implications on operationalized models (model upgrades, package/library updates, version control, etc.).