Module 1: Tableau Foundations

Connecting to Data

Every chart you will ever build in Tableau starts with a data connection. If the connection is wrong, for example the wrong sheet, a header row treated as data, or two tables joined in a way that duplicates rows, every number on your dashboard will be wrong too, however good it looks. Taking ten minutes to connect carefully saves hours of confusion later.

The Tableau products you will meet

Tableau comes in several forms. You only need one of the first two to follow this course.

Product names, editions and licensing change from time to time, so check the current details on Tableau's official website before buying anything.

Making a connection step by step

1. Open Tableau. The Connect pane on the left lists file types (Microsoft Excel, Text file, JSON, PDF, spatial files) and, in Tableau Desktop, servers such as Microsoft SQL Server, MySQL, PostgreSQL and Google Sheets.
2. Choose your file or server and sign in if asked. You land on the Data Source page.
3. Drag a sheet or table from the left into the canvas at the top. Tableau shows a preview grid underneath.
4. Check the preview: are the column headers correct, are numbers right-aligned as numbers, are dates shown as dates? Each column has a small data type icon (Abc for text, # for number, a calendar for date, a globe for geographic). Click the icon to change the type if Tableau guessed wrongly.
5. If the first rows of an Excel sheet contain titles or notes rather than headers, tick Use Data Interpreter. It tries to detect the real table and skip decorative rows. Always review what it did.
6. Click the Sheet 1 tab at the bottom to start building.

Live connections versus extracts

On the Data Source page, at the top right, you can choose Live or Extract.

• A live connection sends queries to the original source every time you change a view. You always see current data, but speed depends on the source system and the network.
• An extract is a snapshot of the data saved in Tableau's own .hyper format. Extracts are usually fast and work offline, but they only change when you refresh them, either manually or on a schedule after publishing.

A simple rule: use extracts for spreadsheets and for slow or busy databases; use live connections when people need up-to-the-minute numbers and the database can handle the load.

Combining tables: relationships, joins and unions

Real analysis often needs more than one table. Tableau offers three ways to combine them:

• Relationships (the default "noodles" you see when you drag a second table onto the canvas) link tables on a common field but keep them separate. Tableau works out the right join for each chart, which avoids many duplication problems.
• Joins (inner, left, right, full outer) physically merge tables into one wider table. You reach them by double-clicking a table to open its physical layer.
• Unions stack tables with the same columns on top of each other, for example January, February and March sales sheets.

Worked example: a Nairobi electronics retailer

Imagine a shop chain with branches in Westlands, CBD and Thika Road. The finance team sends you an Excel workbook with two sheets: Sales (Order ID, Date, Branch, Product ID, Quantity, Amount KES) and Products (Product ID, Product Name, Category, Unit Cost KES). All figures are hypothetical.

1. Connect to the workbook and drag Sales onto the canvas.
2. Drag Products next to it. Tableau proposes a relationship on Product ID. Confirm the field names match on both sides.
3. Check that Date shows a calendar icon and Amount KES shows #. If Amount KES shows Abc, there may be text such as "KES 1,200" in the cells; fix this at the source or with a calculation later.
4. Switch to Extract because the file sits on a shared drive and you want speed.
5. Go to Sheet 1 and drag Amount KES into the view. Compare the total with the total in Excel. If they match, your connection is sound.

Common mistakes to avoid

• Not reconciling totals. Always compare one grand total in Tableau with the source before building anything else.
• Using joins when a relationship is enough. Joining a sales table to a targets table at different levels of detail can repeat rows and inflate sums.
• Forgetting to refresh extracts. An extract made last month shows last month's world.
• Publishing private data to Tableau Public. Anything saved there can be seen by anyone.