Learn
// THE STORY OF DATA
Learn data from scratch.
One chapter at a time.
Every concept is a chapter. Every chapter builds on the last. Pick your learning path below — Data Engineering, Cloud Platforms, or AI & Intelligence.
—
Total chapters
—
Live now
3
Learning paths
∞
Always free
Loading chapters...
ERA 2
Cloud Platforms
▼
CH.1
Snowflake deep dive
Architecture, virtual warehouses, zero-copy cloning, time travel
🔮 ComingCH.2
Databricks deep dive
Unified analytics, Delta Lake, MLflow, Photon engine
🔮 ComingCH.3
Google BigQuery
Serverless, columnar, ML inside SQL, cost optimisation
🔮 ComingCH.4
AWS Redshift & S3
RA3 nodes, Spectrum, data sharing, the AWS data ecosystem
🔮 ComingCH.5
Azure Synapse & ADLS
Dedicated pools, serverless SQL, Purview integration
🔮 ComingCH.6
AWS vs Azure vs GCP
Which cloud for data? Honest comparison for data engineers
🔮 Coming
ERA 2
Engineering Tools
▼
CH.7
dbt — the analytics engineer
Models, tests, docs, lineage — the tool that changed analytics
🔮 ComingCH.8
Apache Spark explained
RDDs, DataFrames, Spark SQL, when to use it vs alternatives
🔮 ComingCH.9
Airflow vs Prefect vs Dagster
Pipeline orchestration — which tool, when, and why
🔮 ComingCH.10
Kafka — real-time streaming
Topics, partitions, consumer groups, Kafka vs Kinesis
🔮 Coming
ERA 3
AI Foundations
▼
CH.1
How LLMs actually work
Transformers, attention, tokens — no PhD required
🔮 ComingCH.2
Vector Databases
Embeddings, similarity search, Pinecone, Weaviate, pgvector
🔮 ComingCH.3
RAG — Retrieval Augmented Generation
How AI finds the right context — the architecture behind smart chatbots
🔮 ComingCH.4
AI Agents explained
What makes an agent, tool use, memory, planning loops
🔮 ComingCH.5
Agentic AI workflows
Multi-agent systems, orchestration, autonomous pipelines
🔮 ComingCH.6
Fine-tuning vs Prompt Engineering
When to fine-tune, when to prompt — the real trade-offs
🔮 ComingCH.7
The AI-native data stack
How AI changes every layer of the data stack
🔮 ComingCH.8
GPT vs Gemini vs Claude
Every version, every benchmark, real-world performance compared
🔮 Coming📖 Business domain knowledge — your competitive edge. These are KPIs and concepts an AI engineer won't know but a data engineer with functional experience will. Use this as a reference before any client meeting or architecture discussion.
📦 Supply Chain & Operations
Open Orders
Orders placed but not yet fulfilled, shipped, or invoiced. The "work in progress" of sales. Varies by industry — in manufacturing includes released, confirmed, and backordered lines.
Open Orders = Orders Received − Orders Shipped
Fill Rate
Percentage of customer demand met from available stock without backorders. Line fill rate vs order fill rate vs case fill rate are all different.
Fill Rate = Units Shipped / Units Ordered × 100
OTIF — On Time In Full
Deliveries arriving on time AND with correct quantity. Walmart fines suppliers for OTIF failures. Gold standard of supply chain performance.
OTIF % = Orders On-Time & In-Full / Total Orders × 100
Inventory Turnover
How many times inventory is sold and replaced in a period. High = lean and efficient. Low = overstocked and cash-heavy.
Turnover = COGS / Average Inventory
Days Sales Outstanding (DSO)
Average days to collect payment after a sale. Measures cash collection efficiency. CFOs watch this weekly.
DSO = (AR / Total Credit Sales) × Days
OEE — Overall Equipment Effectiveness
Gold standard for manufacturing productivity. World class = 85%. Most plants run at 60%.
OEE = Availability × Performance × Quality
💰 Finance & Revenue
Revenue Recognition
When revenue is "earned" vs when cash is received. Under IFRS 15/ASC 606, recognised when performance obligation is met — not when invoiced or paid. Shapes your entire data model.
Gross Margin
Revenue minus COGS. Shows how profitable the core product is before operating expenses.
GM % = (Revenue − COGS) / Revenue × 100
EBITDA
Earnings Before Interest, Tax, Depreciation, Amortisation. Proxy for operational cash flow. Every board deck has it.
Deferred Revenue
Cash received but not yet earned — a liability. Common in SaaS annual subscriptions. An AI engineer asked to "show total revenue" will miss this entirely.
Working Capital
Current assets minus current liabilities. Measures short-term financial health. Negative working capital = potential liquidity crisis.
Working Capital = Current Assets − Current Liabilities
🛒 Sales & CRM
Pipeline Coverage
Ratio of pipeline value to sales target. 3x = $3 in pipeline per $1 of target. Below 3x = alarm bells for sales leaders.
Coverage = Pipeline Value / Sales Target
Customer Lifetime Value (CLV)
Total revenue expected from a customer over entire relationship with the business.
CLV = AOV × Purchase Frequency × Lifespan
Churn Rate
% of customers who stop using your product. Revenue churn vs customer churn are different — know both.
Churn = Lost Customers / Start Customers × 100
Net Revenue Retention (NRR)
Revenue from existing customers including expansions and churn. Over 100% = growing without new customers. Best SaaS health metric.
NRR = (Start MRR + Expansion − Churn) / Start MRR
CAC — Customer Acquisition Cost
Total cost to acquire one new customer. The ratio of CLV:CAC determines business viability. 3:1 is healthy.
CAC = Total Sales & Marketing Spend / New Customers
🏭 Manufacturing & ERP
Bill of Materials (BOM)
Hierarchical list of all components to build a product. Multi-level BOMs are recursive — most AI tools fail at BOM explosion without guidance.
MRP — Material Requirements Planning
Calculates what materials are needed, how much, and when — based on demand forecast and BOM. Drives procurement and production scheduling.
Capacity Utilisation
% of total production capacity being used. Too high = bottlenecks. Too low = wasted fixed costs.
Utilisation = Actual Output / Maximum Capacity × 100
Planned vs Actual
Variance between what was planned and what happened. Foundation of every management dashboard in manufacturing, finance, and sales.