Tableau Desktop: Data Connection and Preparation

6.1 Module 2 kickoff: from theory to hands-on Tableau

6.1.1 Where the course stands and how this module works

Module one is done — that was the theoretical part of the course. Module two starts today, and from here on the sessions are mostly hands-on: a combination of live demos and example sharing, with Tableau Desktop used throughout. The plan for the next few classes is to spend most of the time working inside Tableau itself.

The instructor made a point of preparing much more than the couple of lines the course handout listed for this session; the reasoning is simple — the goal is not to memorize a feature list but to understand the product deeply enough that you can open it and independently create things. A second, explicit goal: by the end of these sessions you should be able to walk into an interview and speak about this tool with confidence. Everything builds toward dashboarding and storytelling, which is where the course is headed step by step.

Because the session is demo-heavy, the instructor warned that the walkthroughs might jump around a little (finding data files, dragging, dropping) and asked for patience. The session was also designed to be re-watchable: the class was told they could always review the video again. The overall teaching stance for the rest of the course: do not limit the content to the handout; expect the sessions to go well beyond it, because the aim is that you can create things on your own afterward.

6.1.2 The Tableau product family

Tableau is not a single tool but a cluster of products, and it helps to know what each one is for, even if you will not use all of them:

• Tableau Desktop — the individual analysis tool. If you are doing something for yourself at an individual level, this is the one. Every demo in this course uses Desktop. It is a complete tool by itself; you do not need the others to work with it.
• Tableau Server — the collaboration layer. You publish your reports and outcome worksheets here, and other people can see them on the server itself.
• Tableau Reader — strictly for reading and viewing. It cannot alter, create, or change any content that Desktop or Server produced. It is a view-only companion.
• Tableau Prep and Tableau Data Management — add-ons used when working on very large programs and huge datasets. From an individual Desktop point of view you do not need them.
• Tableau Mobile — the mobile companion.
• Tableau Public — the free platform anyone can use. It does not come with all the features of Desktop, but it is a legitimate free starting point.
• Tableau Extensions — the ecosystem of add-on extensions.

A useful way to sort this family in your head: Desktop creates, Server shares, Reader views, Prep cleans at scale, Public lets anyone start for free. The product names appear all the time in job postings and interview conversation, so even a one-line sense of each one is valuable.

Real-world: Tableau is owned by Salesforce, and the class included someone with a Salesforce background who was invited to share any real-world Tableau associations from their work during the course. The official website walks through every product and offers a demo for each; that is the best place to browse the full list.

6.1.3 Installing Tableau: trial and student license

There are two ways to get Tableau on your machine. The free trial lasts only 14 days — probably not long enough to work through a course, so treat it as a quick look. The better route for students is the dedicated student page: you enter your details (name, the institute you study at, what you will use it for), and the site emails you a license key valid for one year. You need your student email address for this. One classmate had already applied, and one student had already installed Tableau by the time of this session.

The instructor's strong advice: actually install it rather than just watching. Desktop has many more features to play with than Tableau Public, and the one-year student key is free.

6.1.4 How to practice

The follow-on advice for learning: watching demo videos (there are plenty on YouTube) is not enough — you only really learn this kind of tool by hands-on use. Start small with your own day-to-day data: your expenses, your investments, your portfolio. Create an Excel file, connect to it, drag your features and fields around, and see how the visualization reacts. That small start is how everybody learns the tool.

The practice loop to build this week: pick a tiny personal dataset (your monthly spending works well), connect to it, drag a couple of fields onto the shelves, change the chart type, and ask one question each time ("what happens if I drag this field over here?"). Ten minutes of that beats an hour of passive video watching, because the mistakes you make — and undo — are what stick.

6.1.5 The universal BI workflow

Before touching Tableau, it is worth seeing the shape that every BI tool shares. Whether it is Tableau, Power BI, or Qlik, the workflow has the same three functions:

1. Connect to the data — point the tool at your data source.
2. Drag and drop fields — place fields on the shelves; the visualization builds itself.
3. Save, interact, or publish — keep the result, keep exploring it, or share it.

Every tool does exactly this; what changes between tools is the template, the GUI, and the ease of use. Some tools are niche, some are very light. The instructor had used another BI tool only a few days earlier (the demo shown in the previous class) and it took no time at all to get familiar with it and start dragging and building — evidence that once you know this workflow in one tool, the others come fast.

6.2 Connecting to data: the three entry points

6.2.1 The starting screen

When Tableau opens, the very first screen is a data-source picker with three categories:

• File — connect to a file you have on disk.
• Server — connect directly to a database server.
• Saved data sources — reuse data sources Tableau ships with or that you have saved before.

This screen is the gateway to everything: no data, no visualization. You pick one of the three entry points, point Tableau at your data, and the fields appear ready to use.

There is a live demo of the Excel flow: pick the Excel file, and the sheet's columns arrive as fields (Name, Age, Gender), ready to drag. And a useful habit: Tableau's undo behaves like MS Word undo — you can walk back through everything you just did to return to the starting screen.

6.2.2 File-based connections

The file category covers Microsoft Excel, text files (CSV is the common one), statistical files, and PDF files. These are the formats people actually use in practice, Excel and CSV first, with PDF as a capable extra (more on PDFs in the Data Interpreter section). The flow is the same for any of them: pick the file, the columns appear as fields, and you can immediately start dragging fields to build graphs.

6.2.3 Server-based connections

The server category is where the breadth of Tableau shows: the list contains roughly 77 kinds of connectors, ready to use. The instructor clicked through and picked a few at random: Databricks, MySQL, Google Sheets (it lets you log into your Google account and pull the data from there). On top of the ready-made connectors, Tableau offers JDBC and ODBC connections, which cover any database that supports those standards — which is essentially every mainstream database.

The live demo used MySQL on the instructor's laptop: entering the usual credentials (server address, username, password) connected Tableau directly to the database, and every table in it appeared in the data pane, ready for drag-drop and joins between tables. A practical note from the demo: the connection dialog differs per database. MySQL asks for one set of fields, SQL Server asks for different parameters, and so on. But the knowledge you need is always the same three things: the name of the database, the username, and the password. With those, you can point Tableau at any server.

The two demos cover the extremes of the entry-point spectrum — a plain file on disk and a live database server — yet both end at the same place: a data pane full of fields waiting to be dragged.

6.2.4 Saved data sources

The third category covers predefined data sources that come with Tableau itself — no setup needed. The Superstore sample is the one used all through this session: it brings tables like Order, People, and Return with many fields to play with (dimensions, dates, measures). Saved data sources are the fastest way to start experimenting when you do not have your own data at hand.

6.2.5 When to use file versus server

The choice between a file and a server is a scenario question, and the instructor gave clear decision rules:

Use file-based when:

• The data will not change frequently — a monthly extract, a quarterly report, a one-time analysis.
• You have downloaded the data and you do not need real-time values.
• The data is not sensitive — you will be holding it offline and uploading it where needed.

Use server-based when:

• The data is changing constantly, with users entering records.
• You need real-time graphics that track the current state.
• The data is sensitive — it sits on a password-protected server with protective measures in place.

So the mental shortcut: static, not sensitive, no real-time need → file; changing, sensitive, real-time need → server.

6.3 Live connections versus data extracts

6.3.1 What a live connection is

Every time you connect to a source, Tableau asks which kind of connection you want: live or extract. A live connection is a direct connection to the underlying data. If something in the underlying data changes, your reports change too — the moment you refresh, the new value reflects in the graphics in real time. The underlying source is the single source of truth; the report simply reads from it.

6.3.2 What an extract is and why people use it

A data extract is a copy of the data taken as of a particular time and date. It is a subset — a snapshot you carry with you. Once you are working on an extract, changes in the source data do not affect your graphics at all; you are working offline against the saved copy.

Why would anyone choose a snapshot over live data? The advantages the instructor listed:

• You can download a big dataset and work offline — on a plane, at an airport, anywhere without connectivity.
• It is faster — no live connection overhead, no round trips to the database.
• Performance is better for heavy graphics work.

The flip side: live connections are heavier on performance because the tool constantly deals with real-time data. So the trade-off is freshness and space-safety versus speed and portability.

6.3.3 Worked demo: the same Excel through both modes

The instructor showed the difference end to end with a tiny dummy Excel file.

Setup. The Excel file had three columns — Name, Age, Gender — with nine rows of dummy names (A, B, B, and so on). Tableau was connected live, and the columns arrived exactly as they were in Excel. Then two fields were dragged into the view — Name and Age — producing a simple bar chart of the ages.

Live behavior. In the Excel file, one cell (the row labelled ZZZ) was changed to the text "changing for demo", and the file was saved. Back in Tableau, the chart did not update on its own — but the moment the report was refreshed, the change appeared in the graphic. That is the live behavior: refresh propagates the source change into the report.

Extract behavior. The connection was then switched to extract and the extract was saved. Again the Excel file was edited — this time to "changing second time for the demo" — and saved. Back in the report, refresh changed nothing. The graphics kept the old values, because the extract is the copy you chose to work from, not the Excel file itself.

Back to live. Switching the connection back to live and refreshing finally brought "changing second time for the demo" into the chart. One mode propagates changes, the other does not — that single difference is the whole story.

The instructor's summary line: extract is a copy as of that time and date; live is something which can change — and the refresh test is the way to check which one you are on.

6.3.4 Spotting the mode: database icons

There is a small visual tell you might miss at first glance: a live connection shows one database icon on the data source, while an extract shows two database icons. If you ever forget which mode you are in, look at the icon count before you refresh.

6.4 Data types in Tableau

6.4.1 The six field types and their icons

Whenever you export or connect data, each field carries one of roughly six data types, and Tableau shows the type with an icon at the top of the field, so you can read the schema at a glance:

• String — text, shown with an ABC icon. If the field contains alphabets, it is a string.
• Date — shown with a calendar icon (Order Date and Ship Date in the sample data are dates).
• Date-time — a date with a time component.
• Number — shown with a hash (#) icon. This covers both Number (decimal) and Number (whole).
• Boolean — true/false values, shown as Y and N.
• Geographic — shown with a globe icon; location fields like country and city fall here.

Not every table has every type, but these six are the broad categories you will meet in any database. The icons matter because when a complex table arrives with many fields, the icon is how you immediately recognize what kind of field you are holding.

6.4.2 Changing a field's data type

The type Tableau infers is not final. If you upload data and decide a field would be better as something else — say you would rather keep a numeric-looking field as a string — you can change it: right-click the field, go to change data type, and pick the new type. The instructor showed converting an alphabetic field to boolean, noting that this particular change is not logical for that data, but the mechanism is exactly this. Your call on what the right type is.

6.4.3 Number subtypes and special roles

Numbers come in decimal and whole variants. And geographic data is special: location fields get special geographical roles (country, state, city, postal code and so on) that later power map visualizations. Those roles are covered later in the course; for now, recognize them by the globe icon.

6.5 Data Interpreter: automatic data cleaning

6.5.1 What it is and when it appears

Tableau has some built-in intelligence for deciding on and cleaning up data. Give it data that is not very structured and it applies its own logic to tidy it. This is a feature the instructor flagged as under-shared in most demos — and honestly Power BI is more mature on auto-filtering and auto-correction, but Tableau still has real capability here.

Two things about when the Data Interpreter shows up:

• It appears only when Tableau detects a discrepancy — untidy data where it is struggling: merged cells, nulls, blank fields, gaps. For clean, straightforward, structured data the option does not appear at all, because there is nothing to interpret.
• It may not activate for very small tables (few rows and columns), in which case you clean the data offline and bring in a ready-made clean file.

It is always optional — never imposed on you. You can decline its interpretation and clean the data yourself.

6.5.2 Worked example: merged header cells

The demo file was an Excel sheet whose headers used merged cells: the 2017 header spanned two columns, 2016 spanned two, and 2015 spanned two. When Tableau first read the file, it only read the first row of each merged cell — so each year's second column came in as null. It literally could not see the repeated year.

Toggling Use Data Interpreter fixed it instantly: the interpreter built the logic on its own, recognized the merged headers, and propagated the year values — filling 2017, 2016, and 2015 across the three affected rows. The instructor undid and re-enabled it to show the before/after: undo brought the nulls back, and enabling the feature filled them again. The lesson: when you see nulls that look like they should not be null, the Data Interpreter is the first thing to try.

6.5.3 The review report

Data Interpreter does not act silently. A Review Results report tells you exactly what it did to your data, so you stay in control:

• Red cells = what it decided are column headers.
• Green cells = transactional data.
• Dark green highlights = the specific places where it made changes.

In the demo the review pane flagged 27 such changes. The interpreter can detect and bypass common problems on its own: titles, notes, footnotes, and empty cells. If the report shows two separate tables in one worksheet (separated by row or column gaps), it splits them and hands you two tables to play with instead of one mangled one. And the final decision is yours: if you like the interpretation, proceed; if not — "your interpretation is wrong, I will do my own data myself" — you decline and move on.

6.5.4 Worked example: reading a PDF

PDFs are a special case that the Data Interpreter makes practical. The demo used a PDF report of Amazon stock containing a historical data table.

When you pick a PDF, Tableau asks how much of it to read: one page, all pages, or a selected page. The demo chose all pages, and Tableau came back with a separate table for each page — page 1, table 1; page 2, table 1, and so on. Page 1's table was the interesting one: before interpretation, the month-year column was blank. Enabling the Data Interpreter automatically filled the January, February (and so on) month-year values, removed a first line that was of no use, and turned anything it could not use into nulls. Its own report then explained: this line was treated as a header, this table is more or less clean. Page 2's table came in clean already.

The honest framing from the demo: the interpreter's guess is not necessarily accurate — "not necessarily to be accurate, but one can actually fix it". It is a starting point that saves you from doing the entire cleanup by hand, and you can always correct the result afterward.

6.6 Viewing data before building

6.6.1 The table view

Before building any visualization, look at the data. Every table in the data source can be opened in a table view that shows every field and every value — a real table, not a preview. For the Superstore sample, opening the Order table shows its full set of fields (Row ID, Order ID, Order Date, Ship Date, Ship Mode, Customer ID, Segment, Category, Sub-Category, Product Name, Region, City, Country, Postal Code, Sales, Profit, Discount, Quantity, Profit Ratio, and more) with all values. This is the quick way to know what you are actually working with before you start dragging.

6.6.2 Filtering the view

A full table can be overwhelming. The view supports filtering: you can narrow it down — for example, to a slice of the fields — so you are looking at only what matters. Filtering the view is purely for your inspection; it does not remove fields from the data source.

6.6.3 Hide and unhide fields

Once you have looked at the data, you can decide which fields you want to keep in play. Right-click a field and choose Hide — the field disappears from the data pane. This is useful when the field list is heavy and you do not want everything in your visualization.

The trick is getting it back: use Show Hidden Fields in the data-pane menu, and every hidden field reappears grayed out. Then right-click and Unhide the ones you actually want. So hiding is reversible, and you can pick and choose your working field set freely.

6.7 Column formatting: rename, reset, copy, split

6.7.1 Rename and reset

Field names can be changed two ways: double-click the field name, or use the right-click menu. The demo renamed the ID field to "My ID". Renaming is for your working convenience, but the instructor's aside is worth keeping: a team may well reject ad-hoc naming conventions, and you might forget the original name — so Tableau provides Reset Name, which restores the original name the field had at the source. Rename freely, reset freely; the source name is always one click away.

6.7.2 Copy a column out to Excel

You can copy a column just like in Excel: select the column, press Control+C, and paste into another application with Control+V. The demo pasted a column of IDs into a fresh Excel sheet (using the Copy Value option to paste plain values), and Control+A selects everything in the view. This is a quick way to take any column, bring it somewhere else, and keep it for reference.

6.7.3 Automatic split

Some columns carry two ideas in one cell, and Tableau can split them. Right-click the field and choose Split: Tableau detects where the parts separate and creates two new columns automatically.

The demo used Ship Mode, whose values look like "Second Class", "First Class", "Standard Class". After split, the data source gained Ship Mode (Split 1) containing the first word (Second, First, Standard) and Ship Mode (Split 2) containing "Class". If you do not want the split, delete the new columns — everything is reversible.

6.7.4 Custom split

Automatic split guesses; Custom Split lets you say exactly how. Right-click the field, choose Custom Split, and specify the separator character and how many pieces to make.

The demo used Order ID, whose format is something like CA-2017-12345 — a state code, a year, and a number separated by hyphens. With a hyphen as the separator, Tableau asked whether to split only the first field or all fields. Splitting the first field produced just the "CA" part. Splitting everything produced three columns — Split 1 = CA, Split 2 = the year, Split 3 = the number — each usable in a visualization of its own.

The automatic and custom splits are siblings with one difference: automatic split guesses the separator for you, custom split takes your exact instructions. The rest of the workflow — create the columns, use them, or delete them if unwanted — is identical.

This whole area is what the instructor called data prep: before you move into graphics, you verify the data is understood — look at the tables, join whichever tables you want, decide whether you need all fields, check the data types, split what needs splitting — and only when the data is ready do you step into the visualization mode.

6.8 Sorting fields, not data

6.8.1 Sorting the field list

Whatever Excel file or database you connect, the fields arrive in the source's order — Row ID, Order ID, Order Date, and so on, exactly as the table had them. Sometimes you want the field list sorted for easier navigation. From the data pane menu you can sort the fields A to Z — the demo produced Category, City, Country, Customer ID — or Z to A (descending), which puts the last fields first.

There is an option-set nuance here. The first two options (A to Z, Z to A) are the ones to choose when you are working with one table. When you have multiple tables joined (Order plus People plus Return, for example), the per-table sorting options apply the sorting inside each table, in sequence, in one shot — so each table's fields are sorted within that table. Either way you can always go back to the data source to restore the original order.

6.8.2 Sorting data at the column level

Sorting the field list is not the same as sorting the data. For the data itself, click a column header in the table view: first click sorts ascending, second click descending, and a third click returns to the original order. You can sort the data by any field this way — the demo sorted by Ship Date.

6.8.3 The fields-versus-data distinction

This is exactly the kind of distinction the instructor flagged explicitly — "I am not saying sorting data. Just be clear there. I am talking about sorting the columns." The two features sit in different places and do different jobs.

6.9 The visual interface: a guided tour

6.9.1 Menu bar and toolbar

The menu bar is the top strip: File, Data, Worksheet, and the rest. Below it sits the toolbar, where the instructor was doing undos and formatting throughout the session. The toolbar carries the everyday actions:

• Undo / redo — works like MS Word; walk back through anything.
• Save, new dashboard — create a new dashboard directly from the toolbar.
• Clear sheet — empties the whole worksheet; the view goes blank. Use it when you have made a mess and want to start over.
• Duplicate sheet — copy an existing sheet when you want to tweak one thing but keep most of it.
• Ascending / descending sort — sorting from the top menu.
• Fit options — how the visualization is scaled on the page: entire view (compress everything, big or small, to fit the whole page), standard (default sizing), fit width, or fit height. The demo toggled between entire view and fit height to compress the view.
• Fix axis, highlight control — more advanced icons the course will get to; they are among the most commonly used.

6.9.2 Shelves and cards

Two pieces of Tableau vocabulary you will hear constantly:

• Shelf — the rows and columns areas are called shelves because that is where you keep things: you drag a field there, place it, and the visualization uses it. The instructor's phrasing: "shelf is where we keep things."
• Card — the panels used to format the visuals. The marks card (color, size, shape, label), filters, and pages are cards. This is where you style the marks on your chart.

Next to them is the Show Me panel — the grid of graph types Tableau can offer. Click a thumbnail and Tableau rebuilds the current view as that chart type.

6.9.3 Sheets, dashboards, and stories

Tableau organizes work in three escalating containers:

• Sheet (worksheet) — one individual visualization built from the data. You add new worksheets one at a time.
• Dashboard — a combination of individual sheets. If you have employee data, performance data, and sales data, you build one sheet per view, then place the sheets together on a dashboard. A dashboard is nothing but different parameters clubbed together.
• Story — the presentation layer, built on top of dashboards and sheets. A story is a slide-by-slide navigation: sheet one shows the starting point (say, employee count in 2002), sheet two shows the year-wise growth trajectory, sheet three shows attrition — and you walk the management through the flow like a deck, but with live graphics.

The build order matters: individual sheets first, then dashboards, then stories on top of those. Real-world: the class was asked whether Power BI has a story capability; no one could confirm one. Power BI gives you dashboards, but a playable narrative flow with live graphics appears to be a Tableau strength — though feature sets move, so check current versions before relying on that claim.

6.9.4 Data pane, marks card, status bar, and sheet bar

The tour continues with the remaining regions of the window:

• Data pane (left sidebar) — where the tables and fields live: dimensions on one side, measures on another. Any sets or parameters you create later appear here too. The field icons tell the type at a glance (ABC for text, hash for number, calendar for date).
• Marks card — the card used to build coloring, sizing, and labeling.
• Status bar — the strip showing what the view contains, for example "3 rows by 1 column, sum of profit".
• Sheet bar — the tabs at the bottom holding all your worksheets.
• Main worksheet — the big canvas where you drag-drop fields and the graphic appears.

6.9.5 Pills

A field that you drag onto a shelf is called a pill — the name comes from the shape: a field sitting on a shelf looks like a medicine capsule, which is exactly what a pill is. The term is worth knowing because it is used everywhere in Tableau practice. Two rules of pills:

• A blue pill = a discrete field.
• A green pill = a continuous field.

You will hear people say "put the pill on the shelf" — it simply means placing a field into a shelf to build the visualization.

6.9.6 Naming sheets and workbooks

Every sheet gets a name — whatever you type shows on its bottom tab. In the demo, a view showing sales and profit by category was named "my profit", and that name appeared on the tab. The workbook as a whole gets its name when you do Save As; that name sits on the top bar. Saving is a full save: reopen the workbook and all values, visuals, and hierarchies are restored, with the data source refreshing in the background.

6.9.7 Student questions and answers

A second question compared Tableau with a competing product, so the comparison belongs here too:

6.10 Dimensions and measures

6.10.1 What a dimension is

A dimension is a field that categorizes or segments the data. Typically these are text, dates, and locations — things you classify with, like customer names, product categories, regions, and country names. Dimensions are qualitative and descriptive: they put labels on things and break the data into groups.

Three behaviors follow from that:

• Dimensions decide the grouping of a chart. "Show me sales by category" — the category dimension creates one chunk per category.
• Dimensions do not impact calculations — they are not aggregated. Grouping and sorting happen based on them, but you never sum a dimension. If you try to sum a country name, Tableau refuses — the record count is what exists for a dimension.
• Bar graphs and charts are typically built around dimensions; they provide the bars' categories.

6.10.2 What a measure is

A measure is a field that quantifies something — sales, profit, discount, profit ratio, quantity, percentages. Measures are numerical and are the real values in a chart.

Their key behavior is automatic aggregation. When you bring a table containing names and sales, Tableau knows sales is a measure and shows the sum by default. You are not stuck with sum: the instructor showed the aggregation can be changed to average, median, count, and so on — Tableau offers the options and you choose. Measures are the numbers that enter calculations.

6.10.3 The key differences at a glance

6.10.4 Example: sales by category

The live example made the pair concrete. Placing the Category field (a dimension — its name happened to be "category" too) on the shelf and Sales (a measure) next to it produced one bar per category; switching to sub-categories produced more bars, one per sub-category. In every case the dimension decides the grouping and the measure supplies the number on each bar. Country names and customer names are dimensions; the figures displayed beside them are measures.

6.10.5 Why students mix these up

This is a flagged confusion point: students keep interchanging dimensions and measures. The concept is not complex — descriptive versus quantitative, categorize versus quantify — but the terms come up in almost every Tableau conversation, so the instructor treated the distinction as important and repeated the comparison several times. If you take one vocabulary pair away from this session, take this one. The contrast resurfaces throughout the course in every chart you build.

6.11 Discrete and continuous fields

6.11.1 The two natures of fields

Beyond dimensions and measures, every field also has a nature: discrete or continuous — the same distinction you meet in statistics.

• Continuous means no breakage: an unbroken scale you can slide along. Continuous fields show up green in Tableau.
• Discrete means separate and distinct: individual categories or bins. Discrete fields show up blue.

The color rule is automatic and visible: place fields on a shelf and the pills turn blue for discrete and green for continuous. The instructor's example: a line of a share value per day, changing over time, is always continuous. You can sometimes switch a field's nature, but it depends on the field type — you cannot make a category list continuous in a meaningful way.

6.11.2 Examples and where you will see it

You will meet the pairing constantly: date lines and stock-style series run as continuous greens; category bars run as discrete blues. When the drill-down date hierarchy is used (next section), the discrete view gives you year/quarter/month/day stepping, while the continuous view treats the same dates as an unbroken axis. The course will return to discrete versus continuous in the coming weeks with many examples — this session only established the vocabulary and the color coding.

6.12 Hierarchies: drill up and down

6.12.1 What a hierarchy is

A hierarchy is a logical grouping of fields that naturally go together. Instead of dragging country, region, state, city, and postal code onto a view individually every time, you create one hierarchy, drag it once, and then drill up and down through the levels. Tableau shows hierarchies with a special icon in the data pane, so you can spot them at a glance.

The core advantage: you do not have to drag individual fields each time. One hierarchy is one drag, and roll-up/roll-down is a click. The fields still work individually — dragging a field out of a hierarchy does not destroy it — but the hierarchy is there to save work.

6.12.2 The inbuilt date hierarchy

Dates come with a built-in hierarchy — you do nothing to create it. Drag a date field (say Order Date) and a measure (Profit) onto the view, and Tableau automatically starts at YEAR; the plus sign drills down through QUARTER, then MONTH, then DAY. The demo walked all the way down: year-wise (2018), then quarter-wise sales, then month-wise, then day-wise — and collapsed back up the same way. Whatever granularity the stakeholder wants is one click away.

6.12.3 Building your own hierarchy

User-built hierarchies are simple to create: drag one field onto another in the data pane, and Tableau asks for a hierarchy name. The demo grouped Category and Sub-Category into a hierarchy named "my category" — the two fields are always used back to back, so they belong together.

The demo then showed the payoff. A stakeholder asks for numbers by category — one drag of "my category". The same stakeholder then asks for sub-category — one drill-down click, no re-dragging. Without the hierarchy, each request meant dragging fields individually. The hierarchy is user-created, and right-clicking it and choosing Remove Hierarchy turns the grouped fields back into individual ones whenever you want.

6.12.4 The location hierarchy

The sample data came with a ready-made location hierarchy: country, region, state, city, postal code — fields so closely tied that someone grouped them once. The demo drilled country → region → state → city in one shot (the sample only contains the United States as a country, which is why the demo moved down levels quickly). Any hierarchy can be rolled up or zoomed down the same way.

6.12.5 Adding, reordering, and removing fields

Hierarchies stay flexible:

• Add a field — drag it onto the hierarchy; the demo added Manufacturer to "my category".
• Reorder — drag levels up and down; moving Manufacturer to the top changed the drill order to Manufacturer → Category → Sub-Category.
• Remove — right-click → Remove Hierarchy; every field returns to the individual list.
• Save once, reuse forever — the hierarchy lives in the workbook; reopen it later and it is still there, with the data source refreshed in the background.

The instructor's workflow advice: decide with your analyst team how fields should be clubbed, build the hierarchy once, and never re-drag the same set again. "Once you get the fundamentals clear — why we group, why there is a hierarchy — these things make everything easy. It's not rocket science."

6.13 Sorting in visualizations

6.13.1 Quick sort from the axis

Once a chart exists, sorting is a one-click affair. The demo built sales by category with Sales on the label (the bars read 742k, then 719k, and so on) and sorted directly from the axis: click the axis to sort descending, click again for ascending, and a third click returns to the default order. This axis sort is the quick sort.

6.13.2 Sorting from the menu and from the pill

Two more routes do the same job: sorting from the top menu (the ascending/descending toolbar icons), and sorting directly from the pill on the shelf. Whichever route you take, a clear sort option always exists to reset the view to its original order.

6.14 Grouping data values

6.14.1 Why group

Sometimes the raw categories are too many, or you simply want to compare at a coarser level. Grouping lets you combine data values so they display together as one value. Grouping is purely for the view — the underlying data is untouched — and it gives you comparison power the raw table never had.

6.14.2 Worked example: East and West

The demo used the Region dimension with Sales. With sales shown as labels, each region (East, West, Central, South) had its own bar and its own number. The requirement: "I am more interested in the combined sales of East and West." The moves: multi-select East and West, right-click, choose Group — the two regions merge into one value and the number shown is the combined sum. Ungroup restores the four separate regions.

6.14.3 Creating a group from the data pane

The same operation exists in the data pane: right-click a field, Create → Group, and pick the members. The demo created the group from Region by selecting East and West, naming the group, and getting a brand-new grouped field in the data pane — usable instead of the original dimension anywhere in the workbook.

6.14.4 Editing a group

Groups stay editable. Edit Group opens the member list: add South to the East+West group and the group now covers three regions; remove members the same way; multi-select handles large member lists. If the grouping was a mistake, remove it entirely and the field returns to its original members.

6.14.5 Metros versus the rest of the country

The motivating example for grouping was metro-versus-country comparisons: select a few metro cities (Bangalore, Mumbai), club them together, and group everything else as "rest of the country" — then compare the metro total against the rest. Grouping gives you the freedom to analyze at whatever logical level you need; you are not forced to see the field values exactly as the table stored them.

6.15 Default auto-generated fields

6.15.1 The five fields that come with every data source

Every time you bring data into Tableau, roughly five auto-generated fields appear along with your real fields. They are shown in italics — no other field is italic — which is how you recognize them:

• Measure Names
• Longitude (generated)
• Latitude (generated)
• Number of Records (the instructor also referred to it as the total record count)
• Measure Values

With a single table, the record count is per table; with multiple tables joined, each table gets its own count while the other default fields stay common to the whole data source.

6.15.2 What each one shows

The demo made each field concrete with the Superstore Order table:

• Number of Records — the row count of the table: dragging it onto the label showed 9,944 — the total number of order rows.
• Measure Values — all the measure totals together in one view: the record count (9,944), discount (shown around 1,000), profit, profit ratio (which had no value in the demo), and sales at close to 2 million.
• Measure Names — the names of every numerical field in the table, so the measure values can be labeled with which measure they belong to.

A bonus behavior: double-click any measure value and Tableau automatically builds a graphic from it.

6.15.3 Why bother knowing them

These fields are not something most people use daily — usually you create your own criteria and your own calculations. But they explain where the counts come from when you drag a record count or a measure value, and they show up in italics in every data pane, so recognizing them prevents confusion later.

6.16 Recap and roadmap

6.16.1 What this session covered

The full checklist of what was done in this session:

• Installation — strongly recommended; the one-year student license route.
• The universal BI workflow — connect, analyze, share.
• Three connection types — file-based, server-based, saved data sources, and when to choose each.
• Live versus extract — one database icon vs two; refresh propagates or not; extracts work offline.
• Six data types — string, date, date-time, number, boolean, geographic; changeable per field.
• Data Interpreter — auto-cleaning with a review report; merged cells and PDF reading.
• Viewing data — the table view, filtering, hide/unhide fields.
• Column formatting — rename, reset name, copy a column, split, custom split.
• Sorting fields versus sorting data — the distinction the instructor explicitly flagged.
• The interface tour — menu bar, toolbar, shelves, cards, Show Me, sheets, dashboards, stories, pills.
• Dimensions versus measures — the most repeated vocabulary pair.
• Discrete versus continuous — blue pills and green pills.
• Hierarchies — the inbuilt date hierarchy and building your own.
• Sorting — from the axis, the menu, and the pill.
• Grouping — multi-select groups, data-pane groups, editing and ungrouping.
• Default fields — the five italic auto-generated fields.

6.16.2 What comes next

The next sessions go one level deeper into Tableau: joins between tables, metadata management, calculated fields (your own formulas), parameters, and sets — plus the first real, proper visualizations. The end goal is dashboarding and storytelling, and each session builds one step toward that. The closing advice repeated the central message: start using Tableau now, hands-on, so the coming sessions make sense.

Exam Guidance Summary

No exam-specific guidance was given in this session — the class is in the hands-on phase, and the instructor did not mention distributions, question patterns, or exam dates. The strongest signals about what will be expected:

• Terminology is the exam currency of this module. The instructor repeatedly said the vocabulary — shelves, cards, pills, dimensions, measures, live vs extract, hierarchy, grouping — "will help you a long way" and appears in every Tableau conversation. Expect to be asked what these terms mean and when to use each feature.
• The dimension/measure distinction is a flagged confusion point. The instructor explicitly noted that students keep interchanging these two and called the concept important, repeating the comparison several times. Know the contrast cold: descriptive vs quantitative, categorize vs quantify, never-aggregated vs auto-aggregated.
• Decision scenarios are likely questions. The lecture repeatedly posed scenario rules: file vs server (static vs real-time, sensitive vs not), live vs extract (freshness vs offline performance). These rules are natural question material.
• Interview readiness is an explicit course goal. The instructor stated that by the end of the sessions you should be able to discuss this tool confidently in an interview — so treat product knowledge and workflow knowledge as examinable.
• Hands-on skill matters. The instructor's repeated advice: install the student version and practice on personal data (expenses, portfolio), because watching videos is not enough to retain this material.

Key Industry Applications

• Tableau (Salesforce) — the product family this module is built on: Desktop, Server, Reader, Prep, Data Management, Mobile, Public, and Extensions. Tableau is a Salesforce company; classmates with Salesforce backgrounds were invited to share real-world Tableau usage.
• Power BI and Qlik — the same connect-drag-publish workflow across tools; Power BI is more mature at auto-filtering/auto-correction; Power BI appears to lack Tableau's story capability (dashboards yes, playable narratives no — verify against current versions).
• Database connectors in practice — roughly 77 built-in connectors including Databricks, MySQL, Oracle, MongoDB, Google Sheets (logs into your Google account), and SQL Server, plus JDBC/ODBC for any other database; the knowledge you need for any of them is database name, username, password.
• Sample data — the Superstore dataset (Order, People, Return tables; 9,944 order records; close to 2 million in sales) ships with Tableau as a saved data source and is the standard practice dataset.
• PDF data extraction — the Data Interpreter read an Amazon stock report PDF (historical data table) and turned it into tables, filling blank month-year values and dropping unusable lines.
• Common file formats — Excel and CSV are the everyday file connections; statistical files and PDFs are also supported.
• Learning by doing in industry — the recommended practice pattern: take personal data (expenses, investments, portfolio), build a small Excel, and drag-drop your way to understanding before touching real work.