01 · What Is Tableau?¶

Tableau is a business intelligence (BI) and data visualization platform: you connect it to data (spreadsheets, databases, cloud warehouses), and it lets you build interactive charts and dashboards by dragging fields onto a canvas rather than writing charting code. This module maps out the product family so the rest of the course makes sense, and introduces the small fictional dataset every later Level 1 lesson builds on.

1. The Tableau product family¶

Tableau is not one product — it's a family of related tools, and knowing which one a task belongs to avoids a lot of confusion later.

Tableau Desktop — the authoring application. You install it on your machine, connect to data, and build worksheets and dashboards ("workbooks", saved as .twb or .twbx files). This is what you'll use for nearly all of Level 1 and Level 2.
Tableau Public — a free version of Desktop-style authoring, with one restriction: every workbook you save is published publicly to public.tableau.com — there is no private/local save. Great for learning and for a public portfolio; not for confidential company data.
Tableau Server — a self-hosted (on your own or your company's servers) platform for publishing, sharing, and scheduling refreshes of workbooks so other people can view and interact with them in a browser without owning a Desktop license.
Tableau Cloud (formerly Tableau Online) — the same publishing/sharing experience as Server, but hosted by Salesforce/Tableau instead of your own infrastructure — no servers to maintain.
Tableau Prep — a separate application for cleaning and reshaping data before it reaches a workbook: renaming/splitting columns, joining tables, fixing inconsistent values. Covered starting in Level 2.

Which one do I need for this course?

Tableau Desktop (or Tableau Public, if you don't have a Desktop license) for everything through Level 2. Server/Cloud concepts are covered conceptually from Level 2 Module 9 onward and in depth in Level 3 — you don't need your own server to follow those lessons, since the workflows and admin concepts are explained step-by-step with what you'd see in the interface.

2. How Tableau "thinks" about data¶

Tableau's core idea is the worksheet: one visualization, built by dragging fields from the Data pane (left side of the screen, listing every column in your connected data) onto shelves — labeled drop targets like Columns, Rows, Color, Size, Label, and Filters — arranged around the canvas. The Marks card controls the shape of the mark itself (bar, line, circle, etc.) and lets you drop fields onto Color/Size/Label/Detail/Tooltip to encode more information into each mark.

Tableau splits every field into one of two roles:

Dimensions — categorical or descriptive data (Region, Product, Customer Name, Order Date). Dimensions slice your view into groups.
Measures — numeric data meant to be aggregated (Sales, Profit, Quantity). Measures get summarized (SUM, AVG, COUNT, etc.) by default.

Tableau also colors fields blue (discrete) or green (continuous) in the Data pane and on shelves — a dimension is usually discrete/blue, a measure is usually continuous/green, but either can be switched (right-click a field on a shelf → Discrete or Continuous), which changes how it renders (headers/labels vs. an axis).

3. The running dataset for this course¶

Every Level 1 lesson builds against the same small fictional dataset: a retail chain called Northwind Retail, tracking orders across three regions. You don't need to load this into Tableau to follow along — the lessons walk through the exact clicks and show you the resulting numbers — but writing it down now means every calculated field and chart result later is something you can check by hand.

Table: Orders

Order ID	Order Date	Region	Category	Product	Sales	Quantity	Profit
1001	2024-01-05	East	Furniture	Desk	1200	2	180
1002	2024-01-12	West	Electronics	Monitor	450	3	90
1003	2024-02-03	East	Office Supplies	Paper	60	10	18
1004	2024-02-20	Central	Furniture	Chair	800	4	96
1005	2024-03-01	West	Electronics	Laptop	2200	2	330
1006	2024-03-15	East	Furniture	Bookcase	950	1	-40
1007	2024-03-22	Central	Office Supplies	Binders	120	8	42
1008	2024-04-02	West	Furniture	Desk	1100	1	150

Eight orders, four columns of dimensions (Region, Category, Product, Order Date) and three measures (Sales, Quantity, Profit) — small enough to total by hand, which is exactly the point: when a later lesson says "SUM(Sales) for the East region is 2210", you can verify it (1200 + 60 + 950 = 2210).

4. Live connection vs. extract — a first look¶

When you connect Tableau to a data source, you choose between:

Live — every action in Tableau sends a fresh query to the underlying data source. You always see current data, but performance depends on that source's speed.
Extract — Tableau pulls a compressed, optimized snapshot (a .hyper file) of the data into the workbook itself. Fast and portable, but the data is a point-in-time snapshot until you manually or automatically refresh it.

This choice is made the moment you connect to data — covered in full in Module 2.

How It Actually Works¶

Every drag onto a shelf is, under the hood, building an abstract query Tableau's VizQL engine will translate into a real query against whatever you're connected to (SQL for a database/extract, MDX for an OLAP cube). You never see this query directly in Level 1, but knowing it exists explains a lot of later behavior:

Dragging Region to Rows and SUM(Sales) to Columns produces roughly the query pattern SELECT Region, SUM(Sales) FROM Orders GROUP BY Region ORDER BY Region — every dimension on a shelf becomes a GROUP BY term, every measure becomes an aggregate in the SELECT list. This is why adding a second dimension (say, Category) doesn't just add a column — it changes the GROUP BY to Region, Category, which subdivides every existing group into smaller ones (Furniture-East, Electronics-East, etc.) rather than adding new independent rows.
Live vs. Extract (Section 4) changes where this generated SQL runs, not whether it exists. On Live, that SELECT ... GROUP BY is sent to your actual database every time you touch a shelf. On Extract, the same query runs against Tableau's own embedded Hyper engine — a columnar analytical database bundled into the .hyper file — which is why extracts are often faster: Hyper stores each column contiguously and compressed, so SUM(Sales) only has to scan the Sales column, not full rows.
This is also the mechanical reason dimensions and measures behave differently: a dimension is, by construction, something VizQL can put in a GROUP BY/axis-header list, while a measure is something it can only put inside an aggregate function. Manually switching a field to Discrete/Continuous (Section 2) is really telling VizQL whether to treat its values as GROUP BY buckets (discrete, headers) or as a numeric axis domain (continuous, a line/scale) — the same underlying column, two different query shapes.
Hand-verification tie-in: when Module 3 has you check that the East bar reads 2210, what you're really confirming is that VizQL's generated GROUP BY Region query summed the right three rows (1200 + 60 + 950) from the Orders table above — the same arithmetic a SUM(Sales) WHERE Region = 'East' SQL query would perform against a real database.

Cheat sheet¶

Term	Meaning
Tableau Desktop	Authoring app — builds workbooks (.twb/.twbx)
Tableau Public	Free authoring, but every save publishes publicly
Tableau Server	Self-hosted publishing/sharing platform
Tableau Cloud	Vendor-hosted publishing/sharing platform
Tableau Prep	Separate app for cleaning/reshaping data pre-analysis
Dimension	Categorical field (slices the view); default blue/discrete
Measure	Numeric field (gets aggregated); default green/continuous
Data pane	Left-side list of every field in the connected data
Shelf	Drop target for fields (Columns, Rows, Filters, etc.)
Marks card	Controls mark type + Color/Size/Label/Detail/Tooltip
Live connection	Queries the source fresh every time
Extract	Snapshot of data saved into the workbook (.hyper)

Exercise¶

Without opening Tableau yet: using the Orders table above, compute by hand (1) total Sales for each Region, (2) total Profit for the Furniture category across all regions, and (3) which single order has the worst Profit. Write down your three answers — Module 3 will have you build a Tableau view that reproduces the first of these, and you'll be able to check your worksheet against the number you just calculated.