Connecting to Your Data

Dana coordinates operations for a chain of coffee shops, and her numbers live in four places. Daily sales come out of the point-of-sale system as a CSV she downloads each morning. The staff schedule sits in an Excel file on her laptop. The loyalty list lives in a company database that IT manages. And the team tracks supplier orders in a shared Google Sheet. Every week she stitches it into one spreadsheet by hand, and every week something is stale before she finishes.

Tableau exists to end that weekly scramble. Before it can draw a single chart, though, it has to reach your data where it already lives. That reaching is called connecting, and this chapter is about doing it well.

Why this matters

A chart is only as good as the data behind it. Connect to the wrong file, or to a copy that is two weeks stale, and every view you build on top inherits the mistake. Getting the connection right is the quiet foundation under everything else.

The good news is that Tableau can talk to Excel, CSV files, databases, cloud services, and much more, and the pattern is almost always the same. Pick a source, point Tableau at it, preview what is there, and shape it before you build. Learn that pattern once and you can connect to nearly anything.

One definition before we start. A connector is the built-in adapter Tableau uses to speak a particular source’s language. There is one for Excel, one for SQL Server, one for Google Drive, and so on. You do not build connectors; you choose the right one from a list.

The Connect pane: To a File or To a Server

When you open Tableau Desktop, the app you install to build views, you land on the Start page. Down the left side is the Connect pane, your front door to data.

The Connect pane splits your choices into groups. The two that matter most are To a File and To a Server. Once you have connected to something and saved it, a Saved Data Sources group appears below them, holding shortcuts you can reopen with one click.

To a File is for data saved as a file on your computer or a shared drive. The named options include Microsoft Excel, Text file (which covers CSV), JSON file, Microsoft Access, PDF, Spatial file (for maps), and statistical files from tools like SAS, SPSS, and R.

Two of these are worth a plain word each:

Text file is Tableau’s name for the CSV connector. A CSV is a plain text file where each row is a line and commas separate the values. Most business systems can export one, so it is your universal fallback. Dana’s daily point-of-sale download is a job for this.
JSON file holds data in a nested, structured format that many web systems and apps produce. When you connect, Tableau reads the first 10,000 rows to work out the file’s shape, then lays it out as a flat table you can use.

To a Server is for data that lives in a database or an online service, usually run by an IT team or a cloud provider rather than saved as a file. A database is a system built to store large amounts of structured data reliably. The list names common ones such as Microsoft SQL Server, Oracle, MySQL, Amazon Redshift, and Google BigQuery, alongside cloud sources like Google Drive. This is how Dana reaches the loyalty list IT maintains.

The short list you see under To a Server changes over time, since Tableau surfaces the servers you actually use. To see everything Tableau can connect to, select More… at the bottom of the group. That opens the full catalog of connectors.

A note on Google Sheets, since it comes up constantly. Tableau reaches Google Sheets data through the Google Drive connector under To a Server. You sign in with your Google account and select Allow so Tableau can read your files, then pick the sheet you want. The old standalone Google Sheets connector was retired in 2023, so Google Drive is the current path. That covers Dana’s shared supplier sheet.

The Data Source page: your staging area

After you pick a source and point Tableau at it, you land on the Data Source page. This is where you preview your data and tidy it before you ever build a chart.

The page has three main areas worth knowing:

The left pane lists the connection itself: the file and its sheets, or the server and its tables.
The canvas is the large area at the top where you drag tables to combine them. You will do that combining work in the next chapter.
The data grid runs across the bottom and shows a preview of the first 1,000 rows of your data, laid out in columns. This is where you confirm you have the right data, and where you clean up field names and types. A metadata grid view is also available, which lists your fields as rows for renaming or hiding several at once.

Three quick edits on this page will clean up most raw data.

Rename a field. A field is one column of your data, such as Region or Revenue. If a column arrives with an unclear name like col_3, double-click its name in the data grid and type a clearer one. Good names now mean readable charts later.

Hide a field. If a column is clutter you will never chart, right-click its header and select Hide. It stays in the source but leaves your working list clean. You can always bring it back with Show Hidden Fields later.

Change a data type. Every field has a data type, meaning the kind of value it holds. At the top of each column sits a small icon showing that type. Select the icon to change it. The types include number, text, date, date and time, true/false, and geographic (for places Tableau can map). If a column of dollar amounts comes in as text, Tableau will refuse to add it up until you switch it to a number. Fixing types here prevents that headache downstream.

Try It You will connect Tableau to a simple sales file and preview it. If you have a clean spreadsheet of your own, use that. Otherwise create one: make an Excel file named Cafe Sales.xlsx with a sheet holding the columns Store, Product, Units, and Revenue, and a dozen rows of made-up numbers.

Open Tableau Desktop.

In the Connect pane on the left, under To a File, select Microsoft Excel.

Browse to Cafe Sales.xlsx and select Open.

From the left pane, drag your sheet onto the canvas, or double-click it. The data grid at the bottom fills with a preview of your rows.

Look at the icon above the Revenue column. Confirm it shows a number type. If it shows text, select the icon and choose Number (decimal).

Double-click any unclear column name and rename it to something plain.

What did you notice in the preview? Did every number read as a number, or did anything look like text that should have been a value? Hold that thought. It is exactly the kind of thing the next chapter fixes.

Live or Extract: a live line or a fast snapshot

Look at the upper-right corner of the Data Source page. You will see a choice between Live and Extract. This decides how Tableau gets its numbers, and it is worth understanding.

Live keeps a direct line to the source. Every time you open a view or click a filter, Tableau sends a fresh question to the original data and shows what comes back. Your charts always reflect the source as it stands right now.

Extract makes a fast local snapshot. Tableau copies the data you need into a compressed file it can read very quickly, and your charts read from that copy. The file is saved in the .hyper format, powered by Tableau’s built-in speed engine, and it can hold enormous amounts of data. (You may still see older .tde extract files in the wild. That format is retired, so treat it as legacy only.)

A picture helps.

LIVE
  Source  <--- fresh question every time --->  [ Tableau holds only the connection ]  --->  charts (always current)

EXTRACT
  Source  --- snapshot at extract and each refresh --->  [ Tableau holds a fast .hyper copy ]  --->  charts (fast)

Each choice buys you something and costs you something.

Live is always current, because it reads the source every time. The cost is speed: if the source is slow or far away, your charts wait on it, and you need a working connection whenever you want to look.

Extract is fast, because the data sits in a snapshot built for speed, and it works even when the source is offline. The cost is freshness. A snapshot is only as current as the last time you refreshed it. Refresh means telling Tableau to pull the data again, which you can do by hand or on a schedule.

How to decide, in plain terms:

Choose Extract for most everyday reporting, and when the source is large or slow. A snapshot with a scheduled refresh keeps things quick and steady. This is a safe default when you are learning.
Choose Live when people genuinely need this minute’s numbers, such as a live operations board, and the source is fast enough to keep up.

Dimensions and measures: how Tableau reads your fields

Once your data is in, Tableau sorts every field into one of two buckets in the Data pane, the field list you build charts from. This split shapes every chart you make.

Dimensions are your categories: the things you slice data by. Store name, product, region, and date are dimensions. They answer “by what?” Tableau usually colors them blue.

Measures are your numbers: the things you add up or average. Units sold, revenue, and profit are measures. They answer “how much?” When you drag a measure into a view, Tableau aggregates it, meaning it rolls the individual numbers into a total or an average by default. Tableau usually colors measures green.

Here is the deeper rule behind the colors. Blue marks a field as discrete: Tableau treats each value as a separate labeled bucket. Green marks a field as continuous: Tableau treats it as a flowing range along an axis. Most dimensions are discrete and blue, most measures are continuous and green, which is why the two ideas feel joined. Either can be changed when you need to.

Why does this matter on day one? Tableau’s guesses drive your charts. Drop a green measure on its own and you get a single total. Add a blue dimension and that total splits into a bar per category. If Tableau miscategorized a field, say reading a five-digit ZIP code as a number to sum, your chart goes wrong in a way that looks mysterious until you spot it. A quick glance at the blue and green in the Data pane, fixing anything Tableau got wrong, is a habit that pays off every time.

One layout note. In the Data pane, Tableau separates dimensions from measures with a line, dimensions above and measures below, so you can see its choices at a glance.

In business

Return to Dana. Here is how she rebuilds her week.

The daily point-of-sale download is a CSV, so she connects with the Text file connector. The staff schedule is an Excel file, reached with Microsoft Excel. The loyalty list lives in the company database, so she uses the Microsoft SQL Server connector and signs in once. The shared supplier sheet comes in through Google Drive.

On the Data Source page she renames a few cryptic columns, hides the ones she never charts, and fixes a revenue column that arrived as text. She sets each source to Extract, since none of it needs to be current to the second, and schedules a refresh to run before she arrives each morning. She checks the Data pane and confirms Store and Product landed as blue dimensions while Units and Revenue landed as green measures.

Now the stitching that used to eat her week happens on its own. She opens Tableau and the numbers are already there, current as of the overnight refresh and ready to chart.

Common pitfalls

Loading before previewing. The data grid shows exactly what is inside a source. Building on top of a source you never looked at is how the wrong sheet ends up in a report. Read the preview first.
Connecting to a file that lives only on your laptop. A file saved to your desktop cannot refresh reliably and breaks the moment it is renamed or moved. Put shared data in a shared, permanent location and connect there.
Leaving a number typed as text. If dollars or counts come in as text, Tableau will not add them up. Check the data type icon and fix it on the Data Source page before you build.
Reaching for Live by reflex. “Always current” sounds better than “a snapshot,” so beginners over-choose Live and then wonder why their charts feel slow. For most reporting, an extract with a scheduled refresh is faster and just as useful.
Ignoring blue and green. When a chart behaves strangely, the cause is often a field Tableau typed as the wrong kind. Scan the Data pane and fix miscategorized fields early.

Practice

Connect to two different sources using two different connectors. Try an Excel file and a CSV, or a file and a Google Drive sheet. Preview each in the data grid before you move on.
On the Data Source page, rename one unclear column, hide one column you will never use, and change one field’s data type. Small edits, but they are the start of clean data.
Set one source to Extract and one to Live. Notice where the choice sits in the upper-right corner, and write one sentence on why each setting fits its source.
Open the Data pane and read Tableau’s choices. List which of your fields are blue dimensions and which are green measures. Flag any that look miscategorized.

In the next chapter, you take the data you just connected to and prepare it: combining tables, cleaning messy values, and shaping it so it is ready to chart. Connecting gets the data into the room. Preparing gets it ready to work.

Part of the Tableau track. Developed with AI assistance and reviewed by a human editor. Tableau changes often; menu names, features, and prices may have shifted since the “checked on” dates in this chapter. Verify anything critical against Tableau’s official documentation.