Skip to main content

🎓 Master Curriculum: Data Analytics

From Junior to Senior — A Complete Learning Roadmap​

Scope: Data Analyst (core) + DE literacy (Middle/Senior) + DS bridge (Senior)
Structure: 5 Pillars × 3 Levels (Junior → Middle → Senior)
Philosophy: Tools change, thinking lasts forever.

🗺️ Overview: The 5 Pillars​

PillarVai tròMetaphor
P1 Analytical MindsetTư duy & nền tảng khoa họcBộ não
P2 Data LanguagesNgôn ngữ & kỹ thuật xử lýĐôi tay
P3 Cloud InfrastructureMôi trường & hạ tầng vận hànhNhà xưởng
P4 VisualizationGiao tiếp & trình bày insightGiọng nói
P5 Business & LeadershipTạo ra giá trị kinh doanhTrái tim

Level Definitions​

LevelKinh nghiệmKỳ vọng đầu ra
Junior0–2 nămLàm được task độc lập, cần mentor review
Middle2–5 nămTự dẫn dắt project, bridge DA ↔ DE/DS
Senior5+ nămDefine strategy, lead team, drive culture

🧠 PILLAR 1: Analytical Mindset & Foundations

Gốc rễ của nghề. Công cụ thay đổi, tư duy tồn tại mãi mãi.

Junior Level​

1.1 Types of Analytics — 4 Cấp độ Phân tích​

TypeCâu hỏiVí dụ thực tế
DescriptiveChuyện gì đã xảy ra?Doanh thu tháng 5 là bao nhiêu?
DiagnosticTại sao nó xảy ra?Tại sao doanh thu tháng 5 giảm 15%?
PredictiveChuyện gì sắp xảy ra?Tháng 6 doanh thu có khả năng bao nhiêu?
PrescriptiveChúng ta nên làm gì?Nên tăng ngân sách kênh nào để tối ưu?

1.2 Descriptive Statistics — Thống kê Mô tả​

  • Measures of central tendency: Mean, Median, Mode
  • Measures of spread: Variance, Standard Deviation, IQR, Percentiles
  • Distribution: Normal, Skewed (left/right), Bimodal
  • Outlier detection: Z-score, IQR method

1.3 Probability Basics — Xác suất Nền tảng​

  • Xác suất cơ bản, Xác suất có điều kiện
  • Bayes' Theorem (hiểu khái niệm, không cần tính tay nâng cao)
  • Expected Value

1.4 Cognitive Biases — Thiên kiến Nhận thức​

  • Survivorship Bias — chỉ nhìn thấy "người sống sót"
  • Confirmation Bias — tìm bằng chứng ủng hộ quan điểm sẵn có
  • Sampling Bias — mẫu không đại diện cho tổng thể
  • Simpson's Paradox — xu hướng đảo ngược khi gộp nhóm

1.5 Problem Framing — Đặt vấn đề đúng​

  • SMART Questions framework
  • 5 Whys technique

Middle Level​

1.6 Hypothesis Testing — Kiểm định Giả thuyết​

  • Null hypothesis (H₀) vs Alternative hypothesis (H₁)
  • p-value, significance level (α), statistical power (β)
  • Type I Error (False Positive) vs Type II Error (False Negative)
  • t-test (one-sample, two-sample, paired)
  • Chi-square test (categorical data)
  • ANOVA (so sánh nhiều nhóm)

1.7 A/B Testing Statistics​

  • Sample size calculation (effect size, power analysis)
  • Multiple testing problem (Bonferroni correction)
  • Sequential testing vs fixed-horizon testing
  • Peeking problem (early stopping bias)

1.8 Correlation vs Causation​

  • Pearson, Spearman correlation
  • Confounding variables
  • Spurious correlations
  • Introduction to causal diagrams (DAGs)

1.9 Root Cause Analysis — Phân tích Nguyên nhân Gốc rễ​

  • Fishbone Diagram (Ishikawa)
  • Pareto Analysis (80/20 rule)
  • Drill-down analysis methodology
  • First-Principle Thinking

Senior Level​

1.10 Causal Inference — Suy luận Nhân quả​

  • Difference-in-Differences (DiD)
  • Regression Discontinuity Design (RDD)
  • Instrumental Variables (IV)
  • Synthetic Control Method
  • Causal Impact (Google's framework)

1.11 Advanced Experimental Design​

  • Factorial design & interaction effects
  • Multi-armed bandit algorithms (Thompson Sampling, UCB)
  • Quasi-experiments
  • Network effects in experiments (interference/SUTVA violations)

1.12 Time Series Analysis​

  • Decomposition (Trend, Seasonality, Residual)
  • Stationarity & ADF test
  • ARIMA, SARIMA
  • Facebook Prophet cho business forecasting

1.13 Bayesian Thinking​

  • Bayesian vs Frequentist mindset
  • Prior, Likelihood, Posterior
  • Bayesian A/B testing
  • Credible intervals vs Confidence intervals

⚙️ PILLAR 2: Data Languages & Engineering Literacy

Ngôn ngữ nói chuyện với dữ liệu và tư duy kỹ nghệ phần mềm.

Junior Level​

2.1 SQL — Foundations​

  • SELECT, FROM, WHERE, ORDER BY, LIMIT
  • JOIN types: INNER, LEFT, RIGHT, FULL OUTER, CROSS
  • GROUP BY, HAVING, aggregate functions (COUNT, SUM, AVG, MIN, MAX)
  • Subqueries (correlated vs non-correlated)
  • UNION vs UNION ALL

2.2 SQL — Intermediate​

  • Window Functions: ROW_NUMBER, RANK, DENSE_RANK, LAG, LEAD, FIRST_VALUE, LAST_VALUE
  • Window clauses: PARTITION BY, ORDER BY, ROWS/RANGE BETWEEN
  • CTEs: WITH clause, chaining multiple CTEs
  • CASE WHEN expressions
  • Date/time functions
  • String functions & pattern matching (LIKE, REGEXP)
  • NULL handling (COALESCE, NULLIF, IS NULL)

2.3 Python — Basics​

  • Data types: int, float, str, bool, list, dict, tuple, set
  • Control flow: if/else, for/while loops
  • Functions, lambda functions
  • List comprehensions
  • Error handling: try/except
  • File I/O: đọc/ghi CSV, JSON

2.4 Python — Pandas​

  • Series và DataFrame
  • Data loading: read_csv, read_excel, read_json
  • Data exploration: head, tail, info, describe, shape, dtypes
  • Data selection: loc, iloc, boolean indexing
  • Data cleaning: dropna, fillna, drop_duplicates, replace
  • Data transformation: apply, map, groupby + agg
  • Merging: merge, join, concat
  • Reshaping: pivot_table, melt, stack/unstack

2.5 Python — NumPy​

  • Array creation & operations
  • Broadcasting
  • Statistical functions

2.6 Data Formats & Types​

  • Structured vs Semi-structured vs Unstructured data
  • CSV, JSON, Parquet, Avro — khi nào dùng cái nào
  • Schema design basics

2.7 Git — Basics​

  • git init, clone, add, commit, push, pull
  • Branching: branch, checkout, merge
  • .gitignore
  • GitHub: Repository, Issues, Pull Requests cơ bản

Middle Level​

2.8 SQL — Advanced​

  • Query optimization: EXPLAIN, EXPLAIN ANALYZE
  • Execution plans & query cost
  • Index strategies (B-tree, Hash, Composite index)
  • Recursive CTEs
  • Materialized Views vs Regular Views
  • Stored Procedures & Functions (hiểu khái niệm)
  • Transactions (ACID properties)

2.9 Python — Advanced​

  • Advanced Pandas: MultiIndex, memory optimization, categorical dtype
  • Performance: vectorization thay vì loops
  • Polars (alternative hiệu năng cao hơn Pandas)
  • Regular Expressions với Python
  • Environment management: venv, conda, pyenv

2.10 Python — Visualization​

  • Matplotlib: figure, axes, subplots
  • Seaborn: statistical plots
  • Plotly: interactive charts
  • Streamlit: quick data apps & prototypes

2.11 Data Modeling​

  • Dimensional Modeling (Kimball): Facts vs Dimensions
  • Star Schema: Fact table + Dimension tables
  • Snowflake Schema: Normalized dimensions
  • Grain definition — bước quan trọng nhất trong data modeling
  • SCD Type 1, 2, 3 (Slowly Changing Dimensions)
  • Late-arriving facts & dimensions

2.12 dbt — Foundations​

  • dbt project structure (models, sources, seeds, tests, docs)
  • Materializations: table, view, incremental, ephemeral
  • ref() và source() macros
  • Schema testing: not_null, unique, accepted_values, relationships
  • Documentation: description, meta, tags
  • dbt run, test, docs generate

2.13 ETL vs ELT Patterns​

  • ETL: Extract → Transform → Load (traditional)
  • ELT: Extract → Load → Transform (modern, cloud-native)
  • Trade-offs của từng approach

2.14 API Consumption​

  • REST API concepts (GET, POST, PUT, DELETE)
  • Python requests library
  • Authentication: API Key, OAuth 2.0 basics
  • Pagination & rate limiting
  • Parsing JSON responses thành DataFrame

2.15 Data Quality​

  • Dimensions of data quality: Accuracy, Completeness, Consistency, Timeliness, Uniqueness
  • Data validation rules
  • Anomaly detection basics (Z-score, IQR trên production data)
  • Great Expectations framework

2.16 Git — Intermediate​

  • Branching strategies: GitHub Flow, Gitflow
  • Pull Request workflow & Code Review
  • Merge vs Rebase
  • Cherry-pick, stash, reset

Senior Level​

2.17 dbt — Advanced​

  • Jinja templating & dbt macros
  • dbt packages (dbt-utils, dbt-expectations)
  • Exposures (downstream dependencies)
  • Metrics layer (dbt Semantic Layer)
  • dbt CI/CD với GitHub Actions
  • dbt Mesh & cross-project refs

2.18 Data Governance​

  • Data Lineage: Truy xuất nguồn gốc dữ liệu end-to-end
  • Data Catalog: OpenMetadata, Apache Atlas, Datahub
  • Data Contracts: Schema agreements giữa producer & consumer
  • Data Mesh: Distributed data ownership principles
  • DAMA-DMBOK framework overview
  • GDPR & data privacy compliance

2.19 Orchestration Basics (DE Bridge)​

  • Directed Acyclic Graphs (DAGs) — khái niệm cốt lõi
  • Apache Airflow: DAG structure, operators, sensors, XComs
  • Prefect / Dagster (modern alternatives)
  • Schedule-based vs event-based triggers
  • Monitoring & alerting pipelines

2.20 Batch vs Streaming Architecture​

  • Batch processing: Lambda architecture, khi nào dùng
  • Streaming concepts: event-time vs processing-time, watermarks
  • Kafka basics: topics, partitions, producers, consumers
  • Flink / Spark Streaming awareness
  • Khi nào cần streaming vs batch trong business context

2.21 Testing Strategy for Data​

  • Unit tests cho dbt models
  • Data contract testing
  • End-to-end pipeline testing
  • Monitoring & observability (Monte Carlo, Soda)

☁️ PILLAR 3: Cloud Infrastructure & Modern Data Stack

Nhà xưởng vận hành toàn bộ hệ thống dữ liệu hiện đại.

Junior Level​

3.1 Cloud Fundamentals​

  • Cloud computing models: IaaS, PaaS, SaaS — hiểu khi nào dùng gì
  • Chọn 1 cloud provider để học sâu: AWS / GCP / Azure
  • Regions, Availability Zones, Edge Locations

3.2 Cloud Storage​

  • Object storage: S3 (AWS), GCS (GCP), ADLS Gen2 (Azure)
  • Đọc/ghi data từ cloud storage với Python & SQL
  • Storage classes & lifecycle policies (cost basics)
  • Blob vs File vs Queue storage

3.3 Cloud Data Warehouse — Querying​

  • Google BigQuery: Serverless, columnar, partition/cluster basics
  • Snowflake: Virtual warehouses, databases, schemas
  • Amazon Redshift: Distribution keys, sort keys
  • Microsoft Fabric / Synapse: End-to-end analytics platform
  • Kết nối BI tools vào cloud warehouse

3.4 IAM & Security Basics​

  • Identity & Access Management (IAM)
  • Least privilege principle
  • Service accounts vs user accounts
  • Roles & permissions management

Middle Level​

3.5 Cloud Architecture Patterns​

  • Data Lake: Raw data storage, schema-on-read
  • Data Warehouse: Structured, schema-on-write, optimized for analytics
  • Data Lakehouse: Kết hợp tốt nhất của cả hai (Delta Lake, Iceberg, Hudi)
  • Lambda Architecture vs Kappa Architecture
  • Medallion Architecture (Bronze → Silver → Gold)

3.6 Modern Data Stack Architecture​

Ingestion → Storage → Transform → Semantic → Visualization
(Fivetran)  (BigQuery)  (dbt)    (Metrics Layer) (Power BI)
  • Data ingestion tools: Fivetran, Airbyte, Stitch
  • Reverse ETL: Census, Hightouch (data về operational systems)

3.7 Cloud Data Warehouse — Advanced​

  • BigQuery: Partitioning, Clustering, Materialized Views, BI Engine, Slots
  • Snowflake: Multi-cluster warehouses, Snowpipe, Time Travel, Data Sharing
  • Query cost optimization strategies
  • Caching mechanisms

3.8 Lakehouse Formats​

  • Apache Iceberg: ACID transactions trên data lake, time travel
  • Delta Lake: Databricks ecosystem, Delta tables
  • Apache Hudi: Upserts, incremental processing
  • So sánh: khi nào chọn format nào

3.9 Compute & Serverless​

  • Cloud Functions (Lambda/Cloud Functions/Azure Functions)
  • Containerization basics: Docker cho data engineers
  • Managed Spark: Databricks, EMR, Dataproc
  • Cost model: on-demand vs reserved vs spot

3.10 Microsoft Fabric (End-to-End)​

  • OneLake architecture
  • Lakehouse trong Fabric
  • Data Factory, Dataflow Gen2
  • Semantic Model & Power BI tích hợp

Senior Level​

3.11 Data Platform Engineering​

  • Platform thinking: data platform như một internal product
  • Self-serve analytics platform design
  • Data developer experience (DX)
  • Internal tooling & automation

3.12 Infrastructure as Code (IaC)​

  • Terraform basics: providers, resources, state, modules
  • Managing cloud resources as code
  • CI/CD cho infrastructure

3.13 Streaming Architecture​

  • Apache Kafka: Architecture sâu (replication, consumer groups, compaction)
  • Google Pub/Sub / AWS Kinesis / Azure Event Hubs
  • Stream processing: Flink, Spark Structured Streaming
  • Exactly-once semantics
  • Change Data Capture (CDC): Debezium

3.14 Data Mesh Implementation​

  • Domain-oriented decentralized data ownership
  • Data as a product
  • Self-serve data infrastructure as a platform
  • Federated computational governance
  • Implementing data mesh trên cloud

3.15 Cloud Security & Compliance​

  • Data encryption (at rest, in transit)
  • VPC, Private networking cho data platforms
  • Data residency & sovereignty
  • GDPR, SOC2, ISO27001 awareness
  • Audit logging & monitoring

3.16 FinOps for Data​

  • Cloud cost attribution
  • Reserved capacity vs on-demand optimization
  • Cost anomaly detection
  • Showback/Chargeback models cho data teams
  • BigQuery / Snowflake cost optimization patterns

📊 PILLAR 4: Visualization & Data Storytelling

Dữ liệu chỉ có giá trị khi người khác hiểu được nó.

Junior Level​

4.1 Visual Perception Principles​

  • Pre-attentive attributes: màu sắc, hình dạng, kích thước, vị trí
  • Gestalt principles: proximity, similarity, closure, continuity
  • Working memory limitations (Miller's Law — 7±2)
  • Eye-tracking patterns: F-pattern, Z-pattern

4.2 Chart Selection Guide​

Mục đíchChart phù hợp
So sánh giữa các nhómBar chart, Grouped bar
Xu hướng theo thời gianLine chart, Area chart
Phân phốiHistogram, Box plot, Violin
Tương quanScatter plot, Bubble chart
Tỷ lệ/Cơ cấuPie (≤5 phần), Treemap, Waterfall
Địa lýChoropleth map, Bubble map
Nhiều biếnHeatmap, Parallel coordinates

⚠️ Anti-patterns: Tránh 3D charts, Dual-axis (nếu không cần thiết), Pie chart >5 slices

4.3 Decluttering Principles (Edward Tufte)​

  • Data-ink ratio: maximize data, minimize non-data ink
  • Chartjunk: xóa gridlines thừa, borders, backgrounds
  • Direct labeling thay vì legends khi có thể
  • Consistent color usage

4.4 Color Theory for Data​

  • Categorical palettes: phân biệt nhóm (ColorBrewer)
  • Sequential palettes: low → high (magnitude)
  • Diverging palettes: có điểm trung tính (positive/negative)
  • Accessibility: Color blindness considerations (Deuteranopia)

4.5 BI Tools — Basics (Power BI hoặc Tableau)​

Power BI:

  • Power BI Desktop: Data model, Report, Service
  • Connecting data sources
  • Basic visuals: Bar, Line, Pie, Map, Table, Matrix, Card
  • Filters: Visual-level, Page-level, Report-level
  • Slicers & Cross-filtering

Tableau (alternative):

  • Connecting data, Data pane, Marks card
  • Dimensions vs Measures
  • Basic chart types, Filters, Parameters

4.6 Dashboard Design Principles​

  • Layout hierarchy: most important → top left
  • White space as a design element
  • Consistent color & typography
  • Mobile responsiveness awareness
  • Loading performance considerations

Middle Level​

4.7 Power BI — DAX & Data Modeling​

  • DAX fundamentals: calculated columns vs measures
  • Context: row context vs filter context
  • Core functions: CALCULATE, FILTER, ALL, ALLEXCEPT, RELATED
  • Time intelligence: DATEADD, SAMEPERIODLASTYEAR, DATESYTD
  • Iterator functions: SUMX, AVERAGEX, RANKX
  • Variables trong DAX: VAR...RETURN

4.8 Power BI — Advanced Features​

  • Star schema trong Power BI data model
  • Bidirectional relationships (khi nào nên/không nên)
  • Row Level Security (RLS)
  • Composite models & DirectQuery vs Import
  • Bookmarks, Buttons, Tooltips, Drill-through
  • Performance Analyzer & query diagnostics

4.9 Data Storytelling Framework​

  • SCR Structure: Situation → Complication → Resolution
  • Pyramid Principle: Conclusion first, evidence later
  • Insight vs Finding: "Doanh thu giảm 15%" là Finding, "Vì kênh Facebook CPC tăng 40% làm CAC vượt LTV" là Insight
  • Actionable Insights: mỗi insight phải có recommendation đi kèm

4.10 Executive Reporting​

  • 1-3-10 rule: 1 trang executive summary, 3 trang detail, 10 trang appendix
  • Single-page dashboard: 1 key message rõ ràng
  • KPI cards với context (vs target, vs previous period)
  • Traffic light system cho status reporting

4.11 AI-Assisted BI​

  • Copilot trong Power BI: natural language queries
  • Gemini trong Looker Studio
  • Auto-insights & anomaly detection
  • Q&A visual trong Power BI
  • AI Narratives: automated insight text generation

4.12 Python Visualization (Advanced)​

  • Plotly Dash: interactive web apps for data
  • Streamlit: rapid prototyping cho internal tools
  • Altair: declarative visualization
  • Bokeh: web-ready interactive charts

Senior Level​

4.13 Data Product Thinking​

  • Dashboard-as-a-product: user research cho BI consumers
  • User interviews: hiểu decision-making process của stakeholders
  • Usability testing cho dashboards
  • Product roadmap cho analytics platforms
  • SLA cho data products (freshness, accuracy SLAs)

4.14 Custom & Embedded Analytics​

  • Custom visuals trong Power BI (Power BI SDK)
  • Embedded analytics: Power BI Embedded, Tableau Embedded
  • D3.js basics: SVG, scales, axes (awareness level)
  • Looker Explores & LookML

4.15 Data Journalism Techniques​

  • Annotation-driven storytelling
  • Scrollytelling concepts
  • Information hierarchy for complex narratives
  • When to use static vs interactive

4.16 C-Suite Presentation Skills​

  • Structuring a board-level data presentation
  • Communicating uncertainty to non-technical audience
  • Pre-mortem & scenario analysis presentation
  • Managing tough questions on data methodology

💼 PILLAR 5: Business Acumen, Advanced Analytics & Leadership

Trụ cột tạo ra tiền cho doanh nghiệp và phân biệt DA xuất sắc với DA thông thường.

Junior Level​

5.1 Business Fundamentals​

  • Business model canvas: 9 building blocks
  • Revenue streams & cost structures
  • P&L (Profit & Loss) basics: Revenue, COGS, Gross Margin, EBITDA
  • Unit economics: LTV, CAC, Payback Period

5.2 Universal Business Metrics​

MetricÝ nghĩa
CACCustomer Acquisition Cost
LTV / CLV(Customer) Lifetime Value
Churn RateTỷ lệ rời bỏ khách hàng
NPSNet Promoter Score
GMVGross Merchandise Value
ARR / MRRAnnual/Monthly Recurring Revenue
DAU/MAUDaily/Monthly Active Users
MoM / YoYMonth-over-Month / Year-over-Year growth

5.3 Domain Knowledge — Pick 1 Industry​

Chọn 1 trong các ngành sau để học sâu metric và business model:

  • E-commerce / Retail: Conversion rate, AOV, Cart abandonment, Inventory turnover
  • Finance / Fintech: NPL ratio, Net Interest Margin, RAROC
  • SaaS / Tech: MRR, ARR, Churn, Expansion Revenue, NRR
  • Healthcare: ALOS, Readmission rate, Cost per patient
  • Logistics: OTP (On-Time Performance), Cost per kg, Fill rate

5.4 Requirement Gathering — Basics​

  • Biết đặt câu hỏi TRƯỚC khi phân tích
  • Clarifying questions framework: Who, What, When, Why, How
  • Distinguishing: stakeholder muốn gì ↔ stakeholder thực sự cần gì
  • Documentation: ghi lại assumptions rõ ràng

5.5 Communication — Analysis Writing​

  • Structure: Context → Finding → Implication → Recommendation
  • Quantify everything: tránh "tăng nhiều", dùng "tăng 23% MoM"
  • Know your audience: viết khác nhau cho technical vs non-technical

Middle Level​

5.6 KPIs & OKRs Framework​

  • KPIs vs OKRs: sự khác biệt và cách phối hợp
  • North Star Metric: chọn 1 metric phản ánh core value của product
  • Metric trees: phân rã North Star xuống leading indicators
  • Counter-metrics: tránh optimize 1 chiều
  • Goodhart's Law: "Khi một measure trở thành target, nó không còn là good measure"

5.7 Product Analytics​

  • User Journey Mapping: awareness → acquisition → activation → retention → revenue → referral (AARRR)
  • Funnel Analysis: conversion rate tại mỗi step, drop-off points
  • Cohort Analysis: retention curves, behavioral cohorts
  • Session analysis, pageview analysis
  • Heatmaps & session recordings (Hotjar, FullStory awareness)

5.8 A/B Testing — End-to-End​

  • Hypothesis formulation (cụ thể, measurable)
  • Treatment & control group design
  • Sample size & duration calculation
  • Randomization & assignment strategies
  • Analysis: statistical significance + practical significance
  • Decision framework: khi nào ship, khi nào iterate

5.9 Advanced Business Analysis​

  • RFM Analysis (E-commerce): Recency, Frequency, Monetary
  • Customer Segmentation strategies
  • Market Basket Analysis (Association Rules)
  • Price elasticity analysis
  • Churn prediction fundamentals

5.10 Applied Machine Learning (No-code-from-scratch)​

Supervised Learning:

  • Linear Regression: predicting continuous values
  • Logistic Regression: binary classification
  • Decision Trees & Random Forests: interpretable models
  • Gradient Boosting (XGBoost, LightGBM): production-grade

Unsupervised Learning:

  • K-Means Clustering: customer segmentation
  • DBSCAN: density-based clustering
  • PCA: dimensionality reduction

Practical workflow với scikit-learn:

  • Feature selection & engineering
  • Train/Validation/Test split
  • Cross-validation
  • Model evaluation: MAE, RMSE, R² (regression) / Accuracy, Precision, Recall, F1, AUC-ROC (classification)
  • Hyperparameter tuning: GridSearchCV, RandomizedSearchCV

5.11 Stakeholder Management​

  • Managing up: báo cáo với manager/C-suite
  • Managing sideways: phối hợp với product, engineering, marketing
  • Managing conflict: khi data contradicts stakeholder's belief
  • Expectation setting: timeline, data availability, model limitations
  • "No" với data: cách từ chối request không có giá trị

Senior Level​

5.12 Advanced ML & DS Bridge​

  • Feature engineering nâng cao: encoding strategies, feature interactions
  • Model interpretability: SHAP values, LIME, Partial Dependence Plots
  • MLOps awareness: model versioning, model registry, drift detection
  • Collaborate với Data Scientists: review model design, evaluate business fit
  • Experimentation với ML models: champion/challenger framework

5.13 Causal ML & Advanced Experimentation​

  • Uplift Modeling: identify who benefits from an intervention
  • Treatment Effect Estimation (ATE, ATT, HTE)
  • Causal Impact (Bayesian structural time series)
  • Switchback experiments (cho marketplace/logistics)
  • Long-term effects: handling novelty effects, user learning

5.14 AI/LLM for Analytics​

  • Prompt engineering cho data analysis tasks
  • Text-to-SQL tools: integration & limitations
  • LLM-assisted insight generation
  • Building internal analytics chatbots (awareness)
  • Analyzing unstructured data với LLMs: sentiment, topic modeling, classification
  • Vector databases & semantic search basics

5.15 Data Strategy & Roadmap​

  • Định nghĩa data vision cho tổ chức
  • Data maturity assessment (Gartner Data & Analytics Maturity Model)
  • Build vs Buy decisions cho data tools
  • Data platform roadmap: 6-month, 1-year, 3-year
  • ROI calculation cho data initiatives
  • Building business case cho data investments

5.16 Data Culture & Organization​

  • Types of data teams: centralized, federated, hybrid
  • Embedded analytics model
  • Data literacy programs cho non-technical teams
  • Creating self-serve analytics culture
  • OKRs cho data teams

5.17 Leadership & People Management​

  • Technical mentoring: structured 1:1s, code/analysis reviews
  • Hiring: designing take-home challenges, interview rubrics
  • Onboarding new DA team members
  • Managing performance & career development
  • Cross-functional influence without authority

5.18 Ethics & Responsible Analytics​

  • Algorithmic bias: types, detection, mitigation
  • Fairness metrics cho ML models
  • Privacy by design
  • Data ethics frameworks
  • Communicating model limitations & uncertainty

📋 Curriculum Summary Matrix​

Topic AreaJuniorMiddleSenior
Descriptive Statistics✅✅✅
Hypothesis Testing-✅✅
Causal Inference-🔰✅
SQL Basics & Window Funcs✅✅✅
SQL Advanced & Optimization-✅✅
Python / Pandas✅✅✅
dbt-✅✅ (Advanced)
Data Modeling-✅✅
Data Governance--✅
Cloud Querying✅✅✅
Cloud Architecture-✅✅
Cloud Platform Engineering--✅
Streaming / Kafka-🔰✅
BI Tool (Power BI/Tableau)✅✅✅
DAX / Advanced BI-✅✅
Data Storytelling✅✅✅
Business Metrics✅✅✅
A/B Testing-✅✅
Applied ML-✅✅
Causal ML / LLM--✅
People Leadership--✅

✅ Core requirement | 🔰 Awareness/Intro | - Not required yet

🗓️ Estimated Learning Timeline​

LevelDurationKey Milestones
Junior3–6 monthsSQL Window Functions + Pandas + 1 BI Tool + Dashboard project
Middle6–12 monthsdbt project + Cloud Warehouse + ML model + A/B test end-to-end
Senior12–24 monthsData strategy doc + Lead a cross-functional project + Causal analysis

Made by Anh Tu - Share to be share