Data Visualization

Brittany Fong  | @bfongdata

bfongdata.com

Tableau Ambassador

DC User Group

Why data viz?

Data Viz Tools

PowerBI

A History of Data Visualization

1844 The New York Daily Tribune

1968 New York Times

2012 Five Thirty Eight

2016 The Upshot

1900 Chicago Tribune

2013 Periscopic

1901 The Chicago Tribune

2016 NFL Game Center

2016 Five Thirty Eight

Question

Data
Audience
Creation

Building a data viz

Start with a question

What is the purpose/goal of the visualization?  

What questions are the visualizations going to answer?

How should people feel?

What data do you have?

Is there a unique identifier?

What level of detail is the data at?

Do you need to bring in additional data/tables?

What is the quality of the data?

Cater to your audience

What level of detail do they want to see?

How complex can the visualizations be?

Create a sketch

Pull out a pencil and paper to get the ideas flowing

How will the visualization be viewed? (mobile, print, web)

Common Data Visualization Graphs

Bar Chart

Used for: Comparing categories

Other variations: stacked bar, nested bar, paired bar, multiple bars

Line Graph

Used for: Viewing trends over time

Other variations: year over year trends, add trend line

Scatterplot

Use for: Comparing two measures to see how they relate to each other

Other variations: add trend line, create quadrants

Histogram

Use for: Viewing the distribution of values

                   Ex: Most common number of units sold per order (3 and 2)

Heat Map

Used for: Comparing values by a set color palette

Other Variations: maps

Cartogram

Used for: Geographically mapping a value by color

Less common Data Visualization Graphs

Gantt Chart

Used for: Viewing duration over time

Bump Chart

Used for: Viewing category's rank over time

Other variations: slope chart (comparison from beginning to end)

Unit Chart

Used for: Viewing overall trends in addition to unit level details

Other variations: scatterplot, dot plot, strip plot

Bullet Graph

Used for: Comparing two measures

Other variations: Add indicator colors to background

Data Visualization Best Practices

Why are best practices important?

So we don't end up here!

viz.wtf

Ease of Understanding

Make the visualization as easy as possible to read and understand

Axes

Keep axes clear and consistent

Avoid skipping numbers and Always start from 0

Color

Accessibility - Color Blind (508 Compliant) and make sure your colors have enough contrast (black and white print out)

Clear Labels

Titles, sub-titles, axes, legends, data source, outliers (if necessary)

Try the 30 second test

Context

How do I know if this number is good or bad?  I need some sort of reference.

vs

Keep it simple

Don't add colors when it's not necessary

If it doesn't help explain the visual get rid of it

Don't add too many labels

Check your work

Does the visualization logically make sense?  If you must use a pie chart it better add up to 100%

Use the right graph

Jon Schwabish's Graphic Continuum

What can I do?

Style Guides with Brand compliant color palettes

Example Dashboards

Title Text

Lindsey Poulter    @datavizlinds

Marc Soares    @marc_soares

Chantilly Jaggernauth   @chanjagg

Brittany Fong   

@bfongdata

Data Visualization Resources

Color palettes

pinterest.com

colorbrewer2.org

color.adobe.com

vischeck.com

Data wrangling

Data viz examples and information

• Dear Data (dear-data.com)

• Makeover Monday (http://www.makeovermonday.co.uk/)

• Flowing Data (flowingdata.com)

• The Functional Art (http://www.thefunctionalart.com/)

• Storytelling with Data (http://www.storytellingwithdata.com/gallery/)

• Tableau Public (public.tableau.com)

Data viz resources

• Tableau Website Training Videos

  • (https://www.tableau.com/learn/training)

• Tableau Community (forums, knowledge base, white papers)

  • https://www.tableau.com/learn/training

• Tableau User Groups (local, internal, and subject specific)

  • dctug.com

• Twitter

Visualize your data and influence the world!

Brittany Fong

bfongdata.com

bfongdata@gmail.com