Lecture 5: Advanced Features in Tableau

STA1000 - Data Analytics and Metrics

Damien Dupré

Previously …

  • In Lecture 4 we made figures clear and honest, then prepared the table behind them in Tableau: a pivot, calculated fields, and two files merged into one data source.

Going Forward

In general, through the lectures we have seen variables:

  • Made of character strings (usually interpreted as categorical, but not always)
  • Made of numbers (usually interpreted as continuous, but not always)

But variables can be made of many more values. In business analytics, dates and locations are particularly important: how to handle date and location variables in Tableau?

Tableau is particularly interesting because it creates interactive figures: how to make them smart by customising their interactivity?

Finally, the figures are assembled in a dashboard: how to make them react together, and how to share the dashboard?

1. Date and Location Types in Tableau

Space and Time Variables

So far, we were only interested in two variable types:

  • Categorical (made of characters), also called Dimensions in Tableau
  • Continuous (made of numbers), also called Measures in Tableau

However, we have seen that sometimes…

  • A categorical variable needs to be converted to continuous (e.g., an opinion scale)
  • A continuous variable needs to be converted to categorical (e.g., Survived, coded 0/1, in the Titanic data)

In Tableau, change the type of a variable by drag-and-drop between Dimensions and Measures in the Data pane

In addition to categorical and continuous variables, Tableau has special features to handle Location and Date type variables.

Source: i2tutorials

Date Type Variables

A date refers to a specific time: it can be any combination and format of year, month, day, hour, minute, second …

For example, “March 23, 1984”, “03/23/84”, “23/03/1984” and “1984-03-23” have different formats but refer to the same date.

In data science, we use the ISO 8601 international standard format (YYYY-MM-DD), “1984-03-23”, which sorts dates easily.

By default, a date variable should be properly encoded: a calendar icon appears before it. If it is encoded as categorical (Abc) or continuous (#), convert it to Date or Date & Time. In insta-data.csv, posted_at (2025-01-08 20:00:00) is a Date & Time.

Line and Area Charts

Date variables are usually used as the X axis (Columns in Tableau) to display the evolution of another variable over time.

The resolution of the chart can be adjusted by clicking on the + symbol of the date variable: from the year to the quarter, the month, the day …

Source: Tableau Help

Line and Area Charts

While the date variable is in one field, drag and drop the other variable in the other field. It uses a statistical aggregation: by default the sum (SUM), but it can be changed to count, median, average …

To add a second line in the same figure:

  1. Drag and drop the second variable in Rows, next to the first one
  2. Open the menu of its pill ( right-click / ⌃ Control + click) and select Dual Axis to merge them
  3. If both variables have the same unit (e.g., euros), open the menu of the right axis and select Synchronise Axis

The recordings in this lecture use a Covid-19 sample: the clicks are the same with your own data.

Time as a Dimension or as a Measure

A date can be used as a Dimension (discrete, blue pill) or as a Measure (continuous, green pill). Switch from one to the other in the menu of the date pill: Discrete or Continuous.

As a Dimension: each unit of time (year, quarter, month …) is a category with its own header. The line stops at the end of each year: gaps between the units.

As a Measure: time is a continuous axis, a timeline. The line runs from the first date to the last: no gaps between the units.

A Dimension compares the same units year by year; a Measure follows the trend over the whole period.

Source: Change Date Levels, Tableau Help

Line and Area Charts

More details about Tableau and time series figures, among other resources found online:

Live Example: Loomrow Over Time

Your Turn!

Loomrow over time

With insta-data.csv:

  1. Check that posted_at has a calendar icon
  2. reach by month: which month peaks? Which Lecture 3 post explains it?
  3. spend_eur and attributed_revenue_eur by month on a synchronised dual axis: how much revenue per euro spent?

Location Type Variables

A list of geographical locations by:

  • Continents (Europe, Asia, Africa, Oceania, North America, South America)
  • Countries (Ireland, France, Spain, Germany …)
  • States, regions, counties, cities …

By default, a location variable should be properly encoded: a globe icon appears before it. If it is encoded as categorical (made of characters) or continuous (made of numbers), give it a Geographic Role.

Source: Get Started Mapping with Tableau, Tableau Help

Mapping with Tableau

To create a map:

  1. Drag and drop the location variable in Rows or Columns of the worksheet
  2. Use Show Me and select one of the map figures (the Longitude and Latitude variables are generated automatically):
    • Symbol map: the data are reduced to a point (good for cities)
    • Map: the data fill the full shape of the location
  3. Drag and drop another variable in the Colour box of the Marks card, to display how its values change by location

Source: datadent.org

Mapping with Tableau

More details about Tableau and maps, among other resources found online:

A symbol map (left) and a filled map (right)

A Second File, With Countries

insta-data.csv has no location, but Meta Ads Manager breaks the results of the 127 paid posts down by country: ads-by-country.csv, one row per post and country.

post_id country amount_spent_eur purchases purchase_value_eur
LR0040 Ireland 202.33 8 877.84
LR0040 United Kingdom 59.64 1 104.24
LR0042 Ireland 93.82 5 507.76
LR0042 United Kingdom 44.71 1 96.47
  • Merge it with a relationship on post_id, as in Lecture 4: each table keeps its own rows
  • Check the totals: SUM(amount_spent_eur) is €24,345.78, the spend_eur of insta-data.csv
  • With a join, each paid post appears once per country: SUM(spend_eur) jumps to €106,899.36

Live Example: Mapping Loomrow’s Ads

Your Turn!

Map Loomrow’s ads

From Loop, download ads-by-country.csv, then with insta-data.csv:

  1. Relate the two files on post_id, and check the totals
  2. country should have a globe icon: a symbol map of purchases. Where do the sales come from?
  3. A filled map of the return on ad spend, SUM([Purchase Value Eur]) / SUM([Amount Spent Eur]): where does a euro go furthest?
  4. Belgium and Italy: how many purchases behind them?

2. Interactive Customisations in Tableau

Where We Stopped

Last week, in Tableau:

  • Dates: a Dimension (blue pill, gaps between the units) or a Measure (green pill, one timeline), and two measures on a Dual Axis
  • Locations: a Geographic Role, then a symbol map or a filled map
  • A second file: ads-by-country.csv, related to insta-data.csv on post_id

Today:

  1. Make the figures smart with interactive customisations
  2. Assemble the figures in a dashboard, with actions
  3. The brief of Assignment 1

Aliases and Groups

Variable Alias

Not an interactive feature, but a simple and useful tip: an alias renames the categories displayed by a variable, while the calculations keep the original values.

How: in the Data pane, open the menu of the variable ( right-click / ⌃ Control + click) → Aliases…, then type the new names.

Without aliases: Survived (Titanic data) shows its codes, 0 and 1

With aliases: 0 is displayed as Did not survive, 1 as Survived

Groups

An alias renames one category; a group merges several categories into one: to correct spellings, or to compare families of categories.

How: in the Data pane, open the menu of the variable ( right-click / ⌃ Control + click) → Create → Group…, then:

  1. Select the categories to merge and click Group
  2. Rename the group
  3. Tick Include ‘Other’ to merge all the remaining categories

Shortcut: select marks in the figure, then click the paperclip icon of their tooltip.

Loomrow: campaign mixes three launches (SS25_Launch, AW25_Preorder, SS26_Launch) and three sales (Summer_Sale_25, Black_Friday_25, Deadstock_Drop_26), two groups to compare.

Source: Group Your Data, Tableau Help

Your Turn!

Rename and group

With insta-data.csv:

  1. Alias is_paid: False as Organic, True as Paid
  2. Group the campaigns into Launches and Sales
  3. SUM([Attributed Revenue Eur]) / SUM([Spend Eur]) by campaign group: which group returns more per euro spent?

Marks, Labels and Tooltips

Marks Card

Tableau designs the default figure automatically, according to what it considers the most relevant. The Marks card changes this design:

  1. The type of mark: the drop-down at the top (Automatic, Bar, Line, Circle …)
  2. The design of the marks: Colour, Size, Label, Detail and Tooltip

Each property of the design can be set:

  • For all the marks: click it, e.g., one colour for all the bars
  • By the values of a variable, already used or additional: drag the variable onto it, e.g., format on Colour gives one colour per format, reach on Size bigger marks for a bigger reach

Source: Shelves and Cards Reference, Tableau Help

Labels and Annotations

Labels write the values on the marks. Click Label in the Marks card, then choose the marks to label: all of them, the minimum and maximum, the line ends, or only the selected or highlighted ones.

Annotations point at an insight. Open the menu of a mark ( right-click / ⌃ Control + click) → Annotate, then choose:

  • Mark: attached to a mark, it follows it when the view changes
  • Point: attached to a position in the figure
  • Area: a box around a group of marks

Source: Add Annotations, Tableau Help

Tooltip Principle

The tooltip is a series of information (numeric or graphic) displayed in a temporary box when the cursor goes over, or clicks on, an element of a figure. It gives the actual numbers behind the figure, and tells a short story about what is currently seen.

The default tooltip

A tooltip with conditional formatting

A figure in the tooltip

Sources: InterWorks and Tableau

Tooltip Editing

By default, the tooltip lists the variables of the figure, one per line. Click Tooltip in the Marks card to edit it: the text is free, and each variable between < > shows its value under the cursor (in annotations too).

For example, on reach by month:

Month of Posted At:  <MONTH(Posted At)>
Reach:               <SUM(Reach)>

Will show:

Month of Posted At:  August 2025
Reach:               685,460

Rewritten as a sentence, the same tooltip tells a story:

In <MONTH(Posted At)>, <CNTD(Post Id)> posts reached <SUM(Reach)> accounts!

Will show:

In August 2025, 26 posts reached 685,460 accounts!

Tooltip Advanced

Two ways to put more in a tooltip:

  • Another variable: drag it onto Tooltip in the Marks card, e.g., post_id, counted with Count (Distinct)
  • A full figure: when editing the tooltip, click Insert → Sheets → the name of the figure to add (Viz in Tooltip in Tableau Help)

Your Turn!

Labels and tooltips

On your sheet of reach by month:

  1. Label the highest month, and annotate it with the post that explains it
  2. Rewrite the tooltip as a sentence, with the number of posts
  3. Insert a sheet in the tooltip: reach by format for the month under the cursor

Filters and Highlighters

Local Filters

A filter displays only specific data. Drag a variable onto the Filters shelf, then select what to keep: categories for a categorical variable, values for a continuous one.

Source: Filter Data from Your Views, Tableau Help

Types of Filters

The filter dialog depends on the variable:

  • A Dimension (categories) has four tabs:
    • General: tick the categories to keep
    • Wildcard: the names that contain, start or end with a text
    • Condition: the categories whose measure meets a rule, e.g., a SUM([Reach]) of at least 10,000
    • Top: the top or bottom N, e.g., the 10 posts with the highest SUM([Reach])
  • A Measure keeps a range of values: at least, at most, or between two values
  • A Date keeps a range of dates, relative dates (e.g., the last 3 months), or date parts (e.g., the years 2025 and 2026)

Source: Filter Data from Your Views, Tableau Help

Local Filter Menu

A filter menu (a filter card in Tableau Help) lets the readers make their own selection.

Add it:

  1. Click Analysis in the top toolbar → Filters → the variable the readers choose from
  2. Customise the menu from its drop-down arrow

Note: the same result comes from the Filters shelf: open the menu of the variable → Show Filter.

Choose its style:

  • Single value: a list, a drop-down or a slider
  • Multiple values: a list or a drop-down
  • Range: a slider, for measures and dates

Tick Only Relevant Values to hide the values that the other filters exclude.

Source: Filter Data from Your Views, Tableau Help

One Filter, Several Sheets

By default, a filter only changes its own sheet. On a dashboard, the readers expect one filter menu to change all the figures.

On the Filters shelf, open the menu of the variable → Apply to Worksheets:

  • All Using Related Data Sources: every sheet using insta-data.csv or ads-by-country.csv, related on post_id
  • All Using This Data Source: every sheet using the same data source
  • Selected Worksheets…: the sheets you tick
  • Only This Worksheet: the default

A filter shared by several sheets gets an icon on the Filters shelf.

Source: Apply Filters to Multiple Worksheets, Tableau Help

Local Highlighter Menu

A highlighter menu lets the readers focus the figures on some categories. Unlike a filter, it keeps all the data in the figure: the categories not selected are only dimmed.

Add it:

  1. Click Analysis in the top toolbar → Highlighters → the variable the readers highlight from
  2. Customise the menu from its drop-down arrow (fewer options than a filter menu)

On a dashboard, the highlighter menu applies to all the figures that use the highlighted variable.

Source: Quickly Find Marks in Context with Tableau 10’s New Highlighter, Tableau

Your Turn!

Filter and highlight

  1. A filter menu on format (multiple values, drop-down) on your monthly sheet, applied to all the sheets using insta-data.csv
  2. A bar chart of the 10 posts with the highest reach (a Top filter on post_id): which formats and themes?
  3. reach by theme: a filter on theme, then a highlighter. Select customer in each: where does it rank?

Time Animation Player

Time Animation Player

Drag a date variable onto the Pages shelf to play the figure over time:

  1. A player appears: play forwards or backwards, page by page, at the speed you choose
  2. Tick Show History to keep the marks of the previous pages as trails

Source: Medium

3. Tableau Dashboard

What Is a Dashboard?

For Stephen Few, a dashboard is a visual display of the few things that matter to reach an objective, arranged on one screen and readable at a glance.

Like the dashboard of a car: the few numbers needed to drive, readable in a second, with warning lights when something goes wrong.

Few, S. (2006). Information Dashboard Design: The Effective Visual Communication of Data. O’Reilly.

Tableau Dashboard Showcase

Federica Pinza’s visualisation, DCU winner of the 2024 European Data Visualisation Challenge organised by EDHEC

Tableau Dashboard Showcase

Link to the dashboard

One long page, read from top to bottom:

  • a headline and the key numbers
  • a map, a trend and a correlation
  • the causes, explained with icons and text
  • a conclusion with a call to action

Nice Interactions and Design

Key numbers in colour inside the text, and a Donate button

A year player with Show history, and a highlighter to select a country

Meaningful Visualisations

Two measures on one map, the out-of-school rate (colour) and GDP per capita (circles), with a text to read it

Rural against urban population by region, in the two colours of the dashboard

Insightful Storytelling

The causes: one icon each, short texts, and the key numbers in bold and colour

The conclusion: a summary, and a call to action, raising awareness

Example Not to Follow!

  • Titles: typos, “Wold” and “Countires”
  • Colours: one colour per country on the map, and a pie with one slice per country: the colours carry no information
  • Axes and fonts: raw field names as axis titles (“Obs Value”), two bar charts on different scales, and two different fonts
  • Story: no text to say what the index measures, or what to conclude

Before Building: Purpose and Audience

Tableau’s best practices for dashboard design:

Know your audience

Consider display size

Plan for fast load times

Leverage the sweet spot

Limit views and colours

The first one comes before the first sheet. Answer four questions:

  • Who will read it, and what do they already know?
  • What question does it answer, and what decision will it support?
  • Where will it be read: on a laptop, a phone, a projector?
  • Explore or explain? An exploratory dashboard lets readers search the data; an explanatory one tells them one story. Your assignment is explanatory: it ends with a recommendation.

The 5-second test: after looking at it for 5 seconds, a reader should be able to say what it is about.

Source: Dashboard Design Best Practices, Tableau

Design Strategy

Creating a perfect Tableau dashboard is easy if you follow these steps:

  1. Basic figures: create many figures that can support your story, and keep their default settings
  2. Basic dashboard: put the most relevant figures in a dashboard, with its title, text and pictures, but no design work yet
  3. Figure improvement: go back to the worksheet of each figure and customise it: colour, size, shapes, title, captions, axes and interactions
  4. Dashboard improvement: go back to the dashboard and customise it: background, colours, fonts, layout and actions

Overall Design Guidelines

Keep them in mind at every step:

  1. Don’t overwhelm viewers
  2. Avoid visual clutter
  3. Avoid poor visual design
  4. Don’t add too much data
  5. Provide for consistency
  6. Provide for interaction affordances
  7. Manage complexity
  1. Organise charts
  2. Increase information
  3. Avoid redundancy of information
  4. Design is an iterative process
  5. Context is very important
  6. Use colour carefully

Dashboard Types

Dashboards come in genres: curated ones tell one story (e.g., magazine, infographic), data collections let the readers explore (e.g., analytical, repository). Your assignment is a curated dashboard.

Source: Dashboard Design Patterns

Key Elements

Every dashboard should include these key elements to fill its screenspace:

  • A title
  • A written story, with an introduction, an analysis and a conclusion
  • Multiple visualisations

Optional elements can be added, such as images, buttons or menus.

Example of a dashboard structure, for illustration only: not necessarily to follow.

Key Elements: Title

A good title, as in these real dashboards:

Key Elements: Story

The story, the narrative conveying the information, has three parts:

  • Introduction: presents the context of the issue investigated
  • Analysis: extracts the relevant information from each visualisation, each piece linked to the others (temporal, geographical, logical …)
  • Conclusion: summarises the information and provides recommendations to the reader

Key Elements: Visualisations

  1. Place your visualisations in a relevant order
  2. Make sure they are customised
  3. Add a caption
  4. Add global interactive features

Big Numbers First

Marketing dashboards often open with a few KPIs (Key Performance Indicators) as big numbers: at the top left, each with a comparison. For Loomrow, January to June 2026 against January to June 2025:

2.06M

accounts reached

against 0.45M

6.3%

engagement rate

against 6.4%

€2.75

of revenue per €1 spent

against €3.46

In Tableau: one sheet per number, with the measure on Text in the Marks card, then the sheets side by side in a horizontal container. The engagement rate:

(SUM([Engagements Followers]) + SUM([Engagements Non Followers])) / SUM([Reach])

Layout: the Sweet Spot

Readers scan a screen from the top left, then across and down: put the most important view (or the KPIs) in that corner, and the details below or to the right.

  • Few views: Tableau suggests two or three main ones
  • Align the views on a grid, with white space between the groups
  • Place each legend and filter next to the view it controls
  • Keep the same colour for the same category on every sheet

Source: Best Practices for Effective Dashboards, Tableau Help

Tableau Dashboard

A dashboard is a page made of:

  • Sheets: the figures, each built in its own worksheet
  • Objects: text, images, blank space, buttons …
  • Actions: the interactions that make all the sheets react together

First place the sheets and objects, then change the design features to give the dashboard a professional feel, and add the actions last (step 4 of the design strategy).

Tableau Dashboard

Click the New Dashboard icon at the bottom of the window, then drag the sheets from the left panel onto the page. Tiled keeps them in a tidy grid; Floating lets an object sit on top of the others.

Dashboard Size

Set the size before placing anything, in the Size menu of the Dashboard pane:

  • Fixed size (the default): the same on every screen, with scroll bars if the window is smaller. The safest choice for your assignment, e.g., 1200 × 800 px
  • Range: between a minimum and a maximum size
  • Automatic: fills the window, but the views can be squashed on a small screen

Device Preview shows the dashboard on a desktop, a tablet or a phone, and creates a layout for each device.

Source: Size and Lay Out Your Dashboard, Tableau Help

Layout Containers

A container groups objects that resize together: a Horizontal container shares its width between them (e.g., a row of KPIs), a Vertical one its height (e.g., a column of filters).

Horizontal: side by side

Vertical: one above the other

  • Drag a container from Objects, then drop the sheets inside it; Distribute Evenly gives them the same size
  • Tiled objects share the space without overlapping; a Floating object sits on top, at a fixed position (hold Shift / ⇧ Shift while dragging it)
  • The Layout pane lists every object (the item hierarchy) and sets its position, size, background, border and padding

Source: Size and Lay Out Your Dashboard, Tableau Help

Dashboard Objects

The Objects of the Dashboard pane add everything that is not a sheet:

  • Text: the title, the story, the sources
  • Image: a logo, an icon, a photo
  • Web Page: a web page shown inside the dashboard
  • Blank: space between objects
  • Navigation: a button to another dashboard or sheet
  • Download: a button to export the dashboard as a PDF, a PowerPoint or an image
  • Extension: a feature added by a third party

Any object or container can get a Show/Hide button, e.g., to hide a panel of filters.

Source: Create a Dashboard, Tableau Help

Actions

Basic Actions

Filters and highlighters are set up in each sheet; actions are set up in the dashboard, and make all its figures react together.

Click Dashboard in the top toolbar → Actions → Add Action, then choose:

  1. How to trigger the action: when the cursor goes over a figure (Hover), when it clicks on a figure (Select), or from an added menu (Menu)
  2. Which figure it comes from
  3. Which figures it changes

Start with Filter and Highlight actions: they behave like filters and highlighters, but for the full dashboard. The other actions take more set-up, but they are essential to add buttons, for instance.

Source: Tableau Help

Global Filter and Highlight Actions

With the default settings, an action changes all the figures of the dashboard when the reader hovers over a figure, clicks on it, or selects a value from a menu.

  • Highlight actions are the easiest way to make a dashboard smart: with Hover as the trigger, the figures react together as the cursor moves.

  • Filter actions are useful too, but confusing with Hover or Select as the trigger, as the figures keep changing: trigger them from a Menu instead.

Source: InterWorks

Use as Filter, and Other Actions

The quickest filter action: select a sheet on the dashboard and click its Use as Filter icon (a funnel). A click on one of its marks then filters all the other sheets.

The other actions of the Actions dialog:

  • Go to URL: opens a web page, e.g., the post on Instagram
  • Go to Sheet: moves to another sheet or dashboard
  • Change Parameter and Change Set Values: a click changes a parameter or a set (advanced features)

Source: Actions, Tableau Help

Live Example: Building a Dashboard

Your Turn!

Your Loomrow dashboard

One dashboard, fixed size 1200 × 800 px, to answer: which format should Loomrow post more of, and why?

  1. Three KPIs at the top, in a horizontal container
  2. Four sheets: percent of posts by format, box plots of reach by format, reach by month, return on ad spend by country
  3. A title and a story: introduction, analysis, conclusion
  4. A Highlight action on Hover, and a Filter action on Menu from the map

Start now, finish at home.

Publish

Publishing saves the dashboard online, where everyone can open it.

In Tableau Public in the browser, publishing is saving: click Publish As…, top right.

Warning: everything you publish is public, the data included.

For your assignment, publish your dashboard and submit its URL on Loop.

Before You Publish

A checklist for your dashboard:

Swap with a neighbour and check each other’s dashboard: one compliment, one fix.

Learn More About Dashboards

More details about dashboards in Tableau, among other resources found online:

4. Assignment 1 Brief

Assignment Instructions

Context

You are the marketing analyst at an international company, just hired. Instagram is effectively the label’s whole top of the funnel, and the founders want to know what 18 months of posts tell them.

Task

With the data you will receive by email, tell a story (i.e., a narrative) behind the data in ONE Tableau dashboard, ending with a recommendation.

Submission

Submit the URL of your Tableau Public dashboard on the STA1000 Loop page. Deadline: November 22nd at 12pm midday

Tableau Objective

Obtain the highest mark in Assignment 1

  • Use Tableau Public, and submit the link to a Dashboard (not a Story)
  • Work on each of the four assessment criteria (25% each):
    • Creativity: menus, filters and interactions that do something
    • Analytical depth: figures that answer a question worth asking
    • Design: layout, colour, typography, restraint
    • Overall story: title, narrative, and a conclusion with a recommendation
  • Get inspired by existing dashboards on Tableau Public

Get Inspired

Keep Practising

Thanks for your attention and don’t hesitate to ask if you have any questions!

@damien_dupre

@damien-dupre

https://damien-dupre.github.io

damien.dupre@dcu.ie