[DevDojo] Getting Started with BI: Looker Essentials - 2025

This DevDojo presentation introduces the basics of business intelligence (BI) using Looker and a new approach to data utilization.
It emphasizes the need for companies to gain insights from their vast amounts of data faster and smarter, and redefines traditional BI not as "business reporting" but as data intelligence for making better data-driven decisions.

Transcript

1. Getting Started with BI: Looker Essentials @sathiya BI Engineering, JB Data
2. Thank you Notes • Engineering Growth & Experience Team • Global Operations Team • Tutors • AI
3. サティヤサラティ 　グナセカラン Sathiya Sarathi　Gunasekaran
4. Agenda Slideshow Data Model Creation 02 01 Dashboard Creation 03
5. • Companies - store several PBs of Data (several 1000s of Terabytes) • Derivation of Insights out of data is essential • To conduct Business investigations** ◦ Smarter ◦ Faster ◦ Evidence-based decisions (**Provided the data is clean and easier to consume)
6. • Companies - store several PBs of Data (several 1000s of Terabytes) • Derivation of Insights out of data is essential - How? • To conduct Business investigations** ◦ Smarter ◦ Faster ◦ Evidence-based decisions - How? (**Provided the data is clean and easier to consume) How???
8. Helps answer simple questions like… Business Performance Market Performance Detailed Investigation
9. How is BI, back in the days…
10. BI = Just Business Reports? • BI - Not just for executives tracking KPIs. • In reality, BI is a mindset — using data to make better decisions everywhere. • BI supports not just what happened, but why and what’s next.
11. Match freshness to purpose, not hype Complexity/Cost Data Freshness Fraud detection, logistics tracking Sales reports, Marketing Insights Standard BI setup 🍩 Monthly Daily Hourly Near-Real Time
12. Does it have to be just… …Business Questions? BI isn’t about Business — Better Intelligence… ...❤ powered by Data
13. BI = Insight Automation • Predictive analytics → forecasting demand, churn, or fraud. • Prescriptive insights → recommendations and best next actions. • Embedded analytics → putting insights directly into apps and workflows. So, what do we name it instead?
14. Redefining BI
15. Data intelligence for everyone Function How BI Helps Product Understand feature usage, retention, and user journeys. Engineering Monitor service uptime, latency, and efficiency metrics. Marketing Measure campaign ROI and customer acquisition funnels. Operations Optimize logistics, supply chain, and daily processes. HR/People Track hiring, engagement, and diversity metrics.
16. Fantastic Data Problems and Where to Find them • What is the challenge that you are trying to address? ◦ Your Business problem • Where is your data located? ◦ Locating your datasets and tables • What is the schema? ◦ Your understanding of the schema - the columns • Does it require further transformation? ◦ Any new formulae to be added?
17. 🦥 Traditional flow of Data Operational Datastores Analytical Datastores Insights Data Layer AI/Dashboard Layer SQL🤯 Ref: https://docs.cloud.google.com/looker/docs/viewing-dashboards
18. 🦦 Modern Flow of Data Operational Datastores Analytical Datastores Data Layer Semantic Layer AI/Dashboard Layer SQL High-level Modelling Language Ref: https://docs.cloud.google.com/looker/docs/viewing-dashboards
19. What is a Semantic Layer ? ELEPHANT …What comes to your mind when you see this?
20. What is a Semantic Layer ? If you think of a majestic animal with tusks and a trunk. That’s Semantics — You’re attaching meaning to a symbol.
21. What is a Semantic Layer ? • In Looker, the semantic layer does the same thing: it gives meaning to our raw data. • Instead of tbl_usr_001, we define a User model. • Instead of ord_amt, we define Order Amount. “Analysts don’t have to speak in database language — they can speak in business terms.”
22. Why a Semantic Layer ? • An abstraction layer - To simplify the SQL writing • Translates Data into into Business friendly terms ◦ Business Definition → Database Language • Can accelerate deriving insights from Data • Can dynamically generate Complex SQL with much ease (Programmatically, and not through AI! 🤯) Business Definition Database Language customer_lifetime_value SUM(order_amount) revenue price * quantity
23. Why a Semantic Layer ? Ref: https://cloud.google.com/blog/products/business-intelligence/how-lookers-semantic-layer-enhances-gen-ai-trustworthiness?e=48754805&hl=en
24. Before a Semantic Layer • Raw SQL Everywhere • Analysts and teams write ad-hoc SQLs • Duplication, errors, and conflicting metrics.
25. After a Semantic Layer Let everyone explore — without breaking the warehouse. Looker and LookML are trademarks of Google LLC.
26. Terms that you already know Tables Columns Datasets Projects Ref: https://docs.cloud.google.com/bigquery/public-data#sample_tables
27. Terms that you already know CREATE VIEW AS WITH revenue AS ( SELECT item_categories.name AS category_name, SUM(AMOUNT) AS amount_sold FROM sales INNER JOIN item_categories ON sales.item_category_id = item_categories.id ) SELECT * FROM revenue Views Common Table Expression Aggregations Joins
28. Introducing, Looker Vocabulary Ref: https://docs.cloud.google.com/looker/docs/what-is-lookml
29. LookML - Syntax • A Modelling Language • A simple way to describe your building blocks of data models. • It uses two main building blocks: Parameters and their Values. ◦ parameter: value Fun Fact -- LookML originally was YAML, so it did follow a simpler/more consistent structure. But it was frustrating to parse and edit (for the same reasons YAML is, with the indentation, etc.), and they wanted more flexibility in LookML syntax.
30. Defining a Measure • If you wanted to calculate the sum of the sale_price column, you define a Measure like this: measure: total_revenue { type: sum sql: ${TABLE}.sale_price ;; label: "Total Revenue ($)" } • Giving a user-friendly name (Total Revenue ($)) • Telling Looker the SQL function (sum) to use on the column (sale_price). • Looker writes the complex SQL later! Ref: https://docs.cloud.google.com/looker/docs/reference/param-measure-types
31. LookML Term Analogy Example Syntax view The Table Name view: orders { ... } dimension The Column Name (for grouping) dimension: order_date { ... } measure The Aggregation (for counting/summing) measure: total_sales { ... } sql The SQL Logic sql: ${TABLE}.sale_price ;; type The Function/Type type: sum (for a measure) type: string (for a dimension)
32. Enough Yapping… Show me the Interface
33. Where Business users use the dimensions/measures you created Where you develop the dimensions/measures for Business users to use Access Denied 🤣 Ooh.. Dashboards.. Shiny
34. 2⃣ Let’s check the Develop interface 1⃣ Ooh, flip this switch ON to access the Development Mode
35. Complex SQL with much ease Ooh, flip this switch to access the Development Mode
36. Fields Selected Table & Visualizations Click here for SQL
37. Fields Selected
38. It’s Looker Time ! Let’s explore some Data Models!!