Section outline
-
LC - senior lecturer Pokhodenko B.O. group: Е-20-25
PC - senior lecturer Pokhodenko B.O. group: Е-20-25
Classes are held remotely in the format of web conferences at the link:
On conference Zoomhttps://us04web.zoom.us/j/2272646103?pwd=Y1BGUU4xb2tsYWtVVW13eHdBUEZuUT09
Identification conference: 227 264 6103
Code: 6eqvfm
-
-
Lesson 1. Computer Technologies for Data Processing and Open Data
Objective: formation of students with a holistic system of knowledge, analytical thinking and practical competencies in the field of computer data processing technologies, Data Literacy culture and the use of open government data for effective decision-making. The student must master the phenomenon of data in the digital economy, the ontological concept of DIKW (Data - Information - Knowledge - Wisdom), the concept and structure of Data Literacy, as well as the key stages of data life cycle management (DLM: collection, cleaning/ETL, storage, analysis, visualization, archiving). During the training, special attention is paid to mastering data classification methods (by structuredness and scalar types), studying the engineering framework of data quality metrics (completeness, accuracy, uniqueness, relevance, consistency), as well as studying the regulatory framework of Ukraine in the field of open data (Law "On Access to Public Information", Resolutions of the Cabinet of Ministers No. 835, No. 867) and the infrastructure of the Unified State Web Portal data.gov.ua. Students should learn to analyze and process machine-readable formats (CSV, JSON, XML, XLSX), perform automated collection via CKAN API, clean and normalize datasets, and evaluate the functionality and algorithms of GovTech/B2B services (Opendatabot, YouControl, VKURSI, Clarity App). An important practical result is the formation of the ability to critically evaluate open data, adhere to ethical norms and requirements of the legislation on personal data protection (GDPR), and build effective analytical solutions and dashboards based on objective facts to optimize management and technological processes.
-
Lesson 2. Data Formats, Structures, and Preparation
Objective: formation of students' complex of knowledge, analytical abilities and practical engineering skills in the field of choosing data storage formats, implementing ETL architectural processes, data normalization and cleaning (Data Cleaning), as well as automation of engineering scenarios for preparing information using the No-Code / Low-Code tool Power Query. The student must master the classification of data formats (structured, semi-structured, unstructured), deeply understand the physical, syntactic and comparative specifics of CSV, XLSX, JSON, XML formats, as well as grasp the fundamental concept of ETL (Extract, Transform, Load) / ELT and the GIGO rule (Garbage In, Garbage Out). During the training, special attention is paid to the practical mastering of methods for detecting and eliminating data defects (UTF-8 encoding, gap processing, deduplication, text noise cleaning), mastering the principles of Tidy Data normalization and atomicity of values, as well as studying the procedures for transforming table forms (Pivoting and Unpivoting). An important practical result is the development of the ability to design reproducible deterministic data processing algorithms in the Power Query environment: configure connectors to heterogeneous sources, perform merges (Merge / Append), combine files from folders, and read/edit automatically generated code in the declarative functional language M (Advanced Editor) to create automated data preparation pipelines that are updated in one click without damaging the original source files.
-
Lesson 3. Basic Data Analysis Methods and Tools
Objective: formation of students' complex of theoretical knowledge, critical analytical thinking and practical skills in the field of application of descriptive statistics methods, aggregation, sampling and filtering of data, as well as making informed management decisions based on the concept of Data-Driven Decision Making (DDDM). The student must master the fundamental mathematical concepts of descriptive statistics, the essence of measures of central tendency (arithmetic mean, median, mode) and measures of variability (range, variance, standard deviation), as well as comprehend the structure of the data-driven decision-making cycle and the nature of cognitive interpretation traps. In the learning process, special attention is paid to understanding and comparative analysis of the sensitivity of statistical indicators to outliers (in particular, the robustness of the median), mastering the algorithmic pattern Split-Apply-Combine (split - apply - combine) and studying the architecture and functionality of pivot tables (Pivot Tables). Students should learn to practically apply the tools of analytical spatial and statistical sampling (Sampling), perform simple and multi-criteria logical filtering using AND, OR, NOT, Regex operators and configure multi-level hierarchical sorting of arrays. An important analytical and assessment result is the formation of the ability to differentiate correlation from causation (Correlation vs Causation), identify and eliminate such typical errors of interpretation as Simpson's paradox, survivorship bias, confirmation bias and neglect of the basic scale, which provides a high culture of interpretation of facts and allows you to formulate reliable conclusions for the optimization of engineering and business processes.
-
Lesson 4. Data Visualization and Modern BI Systems. Principles of Dashboarding and Visual Storytelling (Data Storytelling). Overview of Power BI and Google Looker Studio Platforms
Objective: formation of students of a holistic system of knowledge, analytical abilities and practical skills in the field of designing analytical dashboards, applying visual storytelling methods (Data Storytelling) and using modern BI platforms to support management decision-making. The student must master the cognitive and psychological principles of visual perception (attentional attributes, principles of Gestalt psychology), distinguish between exploratory and explanatory visualization, and also comprehend the three-component structure of Data Storytelling (data, visual forms, narrative). During the training, special attention is paid to mastering the methodology and spatial architecture of dashboarding - classification of panels by management levels (strategic, tactical, operational), application of F- and Z-patterns of layout, ensuring a high Data-to-Ink Ratio and compliance with color hygiene. Students should learn to analyze, compare, and select BI tools for specific tasks (distinguishing between the full-featured corporate Microsoft Power BI ecosystem with Power Query, VertiPaq, DAX, and flexible RLS security from the Google Looker Studio cloud-based express reporting service). An important practical and assessment outcome is the ability to identify and eliminate graphic distortions, visual garbage (Chartjunk, 3D effects, truncated Y-axes, misuse of pie charts and dual axes), as well as to eliminate cognitive distortions of perception (Confirmation Bias, Cherry Picking, Base Rate Fallacy, Framing Effect) to create objective, interactive, and ergonomic analytical solutions.
-
Lesson 5. Modern Cloud Databases and Storage Organization. Relational Databases, Tables, Keys. Introduction to Query Language SQL (SELECT, WHERE, JOIN) and Cloud Platforms (Supabase / PostgreSQL)
Objective: formation of students with a complex of fundamental theoretical knowledge, engineering thinking and practical skills in the field of designing relational data models, writing analytical queries in the declarative SQL language and administering modern cloud storage systems (Managed PostgreSQL / Supabase). The student must master the architectural evolution of storage systems from local files to Cloud-Native DBMS, the mathematical foundations of Edgar Codd's relational model (relations, tuples, attributes, domains, 1NF rules), the transactional requirements of the ACID model, and also understand the conceptual differences between row-based transactional OLTP systems and columnar analytical OLAP storages. During the training, special attention is paid to mastering the key system of relational databases - designing primary (PK) and foreign (FK) keys, implementing the principle of referential integrity with cascading actions (ON DELETE CASCADE/RESTRICT), modeling relationships (1:1, 1:N, N:M through intermediate tables), as well as mastering the syntactic and logical order of query execution in the DBMS. Students should learn how to practically compose optimized SQL queries for data selection and projection (SELECT, DISTINCT, calculated fields), apply multi-criteria filtering (WHERE with operators AND, OR, NOT, IN, BETWEEN, LIKE, IS NULL) and perform horizontal joining of sets using JOIN operators (INNER, LEFT, RIGHT, FULL OUTER, CROSS JOIN). An important applied and comparative result is the formation of the ability to evaluate the architectural advantages of managed PostgreSQL (PITR-recovery, Read-replicas, support for hybrid JSONB types and PostGIS/pgvector extensions) and deploy cloud backends on the new generation Supabase platform (using Auto-REST API, Realtime mechanisms via WebSockets, and Row-Level Security (RLS) settings) to create scalable, secure, and high-performance analytical repositories.
-
Lesson 6. Cloud Ecosystems, Collaborative Processing, and No-Code Automation. Google Sheets, Microsoft 365. Foundations of Data Transfer Automation Between Services (Power Automate / Make / n8n)
Objective: formation of students with a complex of theoretical knowledge, architectural thinking and practical skills in the field of organizing collective information processing in modern cloud ecosystems, configuring collaboration tools and designing inter-service integration pipelines based on No-Code / Low-Code platforms. The student must master the prerequisites of evolution from local systems to cloud workspaces, understand the architectural specifics of the Cloud-Native platform of Google Workspace and the hybrid corporate ecosystem of Microsoft 365, as well as grasp the principles of event-driven architecture (Event-Driven Architecture) and the concept of overcoming the problem of "isolated data islands" (Data Silos). During the training, special attention is paid to the practical mastering of collective security and integrity control mechanisms in cloud spreadsheets (separation of Viewer/Commenter/Editor access roles, protection of ranges with calculation formulas, version audit according to the PITR principle and input data validation rules), as well as the use of dynamic cloud functions (IMPORTRANGE, FILTER, UNIQUE, SORT and the SQL-like QUERY function). Students must learn to construct automated data transfer chains, distinguish between the mechanics of polling (Polling) and instant triggers (Webhooks), perform field mapping (Data Mapping), filtering and routing of flows between sources and target actions (Actions). An important analytical and comparative result is the formation of the ability to reasonably choose iPaaS platforms in accordance with the technical and security requirements of the organization (differentiating between the corporate ecosystem with RPA support Microsoft Power Automate, the interactive visual canvas of the Make cloud service, and the Open-Source/Self-Hosted platform n8n with the ability to execute custom code in JavaScript/Python) to create reliable, scalable, and automated analytical pipelines in real time.
-
Lesson 7. Application of Artificial Intelligence for Data Analysis. Copilot, ChatGPT, Google Gemini. Prompt Engineering for Analytics: Formula Generation, CSV Array Interpretation, Ethics, and AI Hallucination Checking
Objective: formation of students with a complex of theoretical knowledge, critical engineering thinking and practical skills in the field of applying generative artificial intelligence technologies for data analysis, development of deterministic analytical prompts, automated interpretation of tabular arrays and ensuring cyber hygiene and verification of results within the concept of AI-Augmented Analytics. The student must master the evolution of analysis paradigms from low-level programming (Pro-Code) to Prompt-Driven analytics, architectural components of augmented analytics (Augmented Data Prep, Discovery, code generation, NLG storytelling), as well as comprehend the probabilistic nature of large language models (LLM) and the mechanisms of the emergence of mathematical and logical hallucinations. During the training, special attention is paid to the comparative analysis of leading cloud ecosystems - distinguishing between the Python Sandbox (Code Interpreter) in ChatGPT, the multimodal environment with a large context window in Google Gemini, and enterprise integration with Power BI and DAX in Microsoft Copilot. Students should learn how to practically design structured queries using the R-C-T-O (Role – Context – Task – Output) framework, apply specialized Chain-of-Thought and Few-Shot Prompting techniques to generate complex SQL queries, QUERY functions, and DAX dimensions, as well as configure the work of autonomous AI agents for step-by-step profiling, cleaning, statistical analysis (EDA), and visualization of CSV arrays. An important analytical and evaluation outcome is the ability to implement a strict four-step anti-hallucination protocol (enforcement of code execution, validation on microsamples, cross-validation, Human-in-the-Loop principle) and ensure compliance with data privacy standards (GDPR, ISO/IEC 27001) through anonymization, PII masking, and the use of secure Enterprise API or Self-Hosted models for safe and informed management decision-making.
-
Lesson 8. Data Cybersecurity, Hygiene, and Prospects (Big Data, AI). Personal Data Protection (GDPR, Ukrainian Legislation), Backup, Two-Factor Authentication. Overview of Data Science, Big Data, and AI Analytics Fields
Objective: formation of a holistic system of knowledge, legal awareness and practical engineering skills in the field of information protection, regulatory compliance, cyber hygiene and orientation in modern technological trends in data engineering. The student must master the evolution of security architectures from outdated perimeter protection to the Zero Trust Architecture (ZTA) model, the fundamental principles of international and national regulatory regulation (GDPR, Laws of Ukraine No. 2297-VI, No. 2163-VIII, ISO/IEC 27001), as well as comprehend the triad of interaction between the areas of Big Data, Data Science and AI-analytics. During the training, special attention is paid to understanding and comparative analysis of the three engineering principles of ZTA (Explicit Verification, Least Privilege Access/JIT/JEA, Assume Breach with micro-segmentation and Data-in-Transit / Data-at-Rest encryption), mastering the concepts of Privacy by Design, Data Minimization and differentiation of the roles of the data controller/processor and the rights of subjects (in particular, the right to be forgotten and access). Students should learn to practically implement personal and corporate cyber hygiene protocols: configure phishing-resistant multi-factor authentication (FIDO2/Passkeys, TOTP), manage credentials through password managers using the Zero-Knowledge model, anonymize and tokenize PII arrays, and design fault-tolerant backup schemes using the extended 3-2-1-1-0 strategy using immutable storage (Immutable Storage / WORM). An important analytical and prognostic result is the formation of the ability to evaluate modern technological vectors of analytics development (real-time streaming processing based on Kafka/Flink, hybrid Data Lakehouse architecture, vector DBMS and RAG systems), as well as to critically analyze and mitigate legal and security risks when deploying corporate analytical pipelines.