Practical Information Visualization

Tableau 1: Introduction
What is Tableau?

State-of-the-art tool to explore data and create interactive information visualization presentations

probably the easiest/fastest-to-use business intelligence & infovis tool out there
✔︎create professional interactive visualizations with zero programming
✔︎integrates easily with many big-data sources out there
✘ does not offer full fine-grained customizability like other toolkits (e.g. programming ones)
✘ certain options are not entirely logical or orthogonal
Quick start guide

Step 1: Select a tool version

•go to http://www.tableau.com/products
Quick start guide

Step 1: Select a tool version

•go to http://www.tableau.com/products
•study versions, pick the one that fits you

Tableau Public
•offers most functionality of other versions
•free of charge and any other usage constraints (e.g. type of user, duration of use, etc)
•cannot save visualizations locally (requires a cloud account, which will also upload data)
•cannot connect to databases (accepts Excel / SPSS / CSV)
•this is the version I use in this course

Tableau Desktop Personal / Professional

•slightly more advanced visualization functionality
•can save visualizations locally (*.twbx format)
•does not require a cloud account (so your data can be kept private)
•options to connect to databases
• Personal: MS Access, OData
• Professional: hundreds of data sources
•15 day free trial
Quick start guide
Tableau Server
•data management options (like a layer atop a basic database)
•data sharing across teams / with users
•author visualizations in Desktop, publish them online from your Server
•the most sophisticated Tableau product (useful for businesses building interactive visualizations
for many clients)

Tableau Online
•web-based front-end for (simple) data visualizations
•basically a Tableau server where you get an account
•can create/view visualizations without downloading/installing tools
•must share your data
•useful for lightweight experimentation
Quick start guide
Step 2: create an online account
•go to https://www.tableau.com/user/login
•create an account (only an e-mail address required, no other info asked)

Here you will

•save your visualizations (and data) built with Tableau Public
•share your visualizations from with other people (e.g. as URLs / embed code)
Quick start guide
Step 3: get a dataset to experiment with
•must be in CSV / Excel / SPSS format
•sample data sources: data.seattle.gov

• public data from city of Seattle

• hundreds of interesting datasets
• governance
• transport
• education
• permits
• safety
• data is entirely free
• anonymous download (no log-in)
• example visualizations present too
Quick start guide
Step 3: get a dataset to experiment with
•let us select the Building Permits dataset
(information on building permits from the last 5 years)

Click here to
examine
dataset
Quick start guide
Step 3: get a dataset to experiment with
•export the data to a local file
•select the Export to CSV option
export data
from here

•we obtain a CSV file of 19.6MB (50K records) (data as of 02/05/2016)

Quick start guide
Step 4: load the data into Tableau
•select the option ‘connect to data / text file’
•load the CSV file

Data Source View

•allows inspecting the data (to see e.g. if loading was OK)
•for now, we don’t use this view further, but go to create a worksheet

create a
worksheet
Quick start guide
Step 5: create a worksheet

Worksheet view
•the place to create visualizations
•typically one worksheet per visualization (chart) select a chart
type here…
•we’ll explain all details later, for now we keep it simple
get your data
from here…

explore/customize
the visualization
here
Quick start guide think of: columns=y, rows=x
..and drop them
Step 5: create our first chart into Columns and
•map of building-permit locations in Seattle Rows respectively

drag Latitude and see a dot these are the averaged

Longitude… appearing here positions of all permits

see video at https://youtu.be/AZ-cy67GJck

Quick start guide
change Latitude and Longitude
Step 5: map of building-permit locations from measure to dimenson

we get now a
dot per record

Explanation
•Tableau aggregates measures (quantitative data), but not dimensions (categorical data)
•changing the field type ‘de-aggregates’ the data in the chart
Quick start guide
Step 5: map of building-permit locations
•dots are too large, they create visual clutter
•change the visual property size of the dots

dots become smaller

change
size here
Quick start guide
Step 6: map of building-permit locations with permit category
•add the Category field to the Color entry in the Marks panel
•dots get now colored by permit-category

a color legend
appears here dot colors =
permit
•note how a color legend is automatically categories
created (since we use color to show data)
Quick start guide
Step 7: map of building-permit locations with permit category and value
•drag the Value measure to the size entry in the Marks panel
•dots are now scaled to show the buildings’ values

a size legend
appears here

Big circles show now expensive building sites (e.g. close to city center)
Quick start guide
Step 8: map of building-permit locations with permit category and value, per given period
•change Issue date field to continuous (so we can select a range of it)
•select ‘Show filter’ for that field filter slider
filter appears appears here
here

We can now show data within a desired time-range of Issue date

Quick start guide
Step 9: how many permits have been issued for each category?
•drag Number of records to Columns and Category to Rows
•this creates a bar chart showing Number of records per Category value

Note
•Number of records is aggregated (since it is a measure)
•Category has five distinct bins (since it is a dimension)
Quick start guide
Step 9: make the bar chart easier to read
•color bars by their Category values
•for this, drag Category to the Color entry in the Marks panel

a color legend
appears here

Categories are now shown in the chart by two things

•textual label
•color
This is called overloading (using several visual variables to show the same data field)
Quick start guide
Step 10: show how the number of permits changed over time
•create a new chart (worksheet)
•drag Number of records to Columns and Issue date to Rows
•we get a time-line (line chart)

Note
•since Issue date is continuous, we get an interpolation (line) between consecutive sample
points (years)
Quick start guide
Step 10: refine the time-scale of the line chart
•select ‘Months’ sampling for Issue date in the Columns entry
•the time-line shows now data at a higher resolution
change sampling of Issue
date to Months here

Note
•observe how fewer permits are issued during the winter months
Quick start guide
Step 11: show permits-issue changes in time per category
•drag Category and drop it to the Color entry in Marks panel
•we get now a set of timelines, one per category

a color legend appears

Note
•see how the permits for single family houses are always at the top, and increasing
Quick start guide
Step 12: show permits-issue changes in time per category and also cumulated values
•change chart type to area chart
•we see now both (changes of) fractions of each permit category and their totals

change chart
type
to area graph
Quick start guide
Step 12: improve readability of area chart
•drag category values in color legend to cluster larger values together
(e.g. put larger values like Single family and Commercial at the bottom)

drag Commercial
here
Putting it All Together
Step 12: create a dashboard to show all results
•select Dashboard panel
•drag and drop our three worksheets in here (map, bar chart, timeline)
Putting it All Together
Step 12: linking visualizations via interactive controls
•for instance: let the Issue date filter affect all three views (not just the map)
•for this, right-click filter and select ‘apply to all worksheets’

Issue date
filter
Putting it All Together
Step 12: linking visualizations via interactive controls (continued)
•let a selection in a view (e.g. the bar chart) act as filter in all other views
•for this, select bar chart, and right-click select ‘use as filter’ select option
‘use as filter’

selecting in here
now filters in the
other views
Putting it All Together
Step 13: adding annotations to a dashboard
•labels, title, icons, hyperlinks
•much like designing a presentation in PowerPoint (albeit, of course, with more limited options)
Putting it All Together
Step 14: saving and sharing
•save using ‘Save to Tableau public…’
•the entire visual design and the data are transferred to Tableau’s cloud server

Share this URL, and

everybody can use
your interactive
visualization

Notes
•the web-based visualization is slower than the desktop one (uses low-speed Tableau server)
•all your data is now public…
Conclusions
Tableau
•state-of-the-art tool to construct information visualization applications
•visualization design, sharing, and interactive use are all covered
•different editions for different user types
•free version available

Visualization options
•all basic charts you know from e.g. Excel
•more advanced charts
• maps, timelines, treemaps
•easy visualization design (no programming required)
•huge design space with a small number of concepts

Next module
•detailed view at Tableau’s options