1 Data Driven Summit 2014
Mark Gamble
Dir Technical Marketing Actuate Corporation
Effective Big Data Visualization
“Every Picture Tells A Story Don’t It?”
2 Data Driven Summit 2014
• What is “data visualization”?
• What is “good”?
• Ed Tufte principles
• Stephen Few principles
• Visualizing data
• Chart basics
• Visualization Examples
• BIRT Style Techniques
Agenda
3 Data Driven Summit 2014
• What is “data visualization”?
• What is “good”?
• Ed Tufte principles
• Stephen Few principles
• Visualizing data
• Chart basics
• Visualization Examples
• BIRT Style Techniques
Agenda
4 Data Driven Summit 2014
What is “data visualization”?
Data Visualization is the “art of information”
It is a depiction of summarized metrics, culled from various sources, combined into a single descriptive graphic.
Data Visualizations are typically employed for quantitative summarization, such as infographics or dashboards
While Big Data presents unique challenges for visualization, the fundamentals of good information design apply
“As big data becomes bigger, and more companies deal with complex datasets with dozens of variables, data visualization will become even more important.”
Julie Steele Editor, Strata - O’Reilly Media
http://radar.oreilly.com/2012/02/why-data-visualization-matters.html
5 Data Driven Summit 2014
What is “good data visualization”?
Different things to different people…..
• Consider your audience
• How will they use the information?
• How “statistically savvy” are they?
• How do they consume the information?
Review the expert opinions
• Ed Tufte
• Stephen Few
6 Data Driven Summit 2014
1. Enforce Visual Comparisons o Draw conclusions faster
o Use thickness, color, weight
o Compare in adjacent space vs over time
2. Show Causality
o Show how one thing makes another occur o Reinforce the meaning of the content
o No point is conveyed without it
3. Show Multivariate Data
o Show data on more than 2 dimensions o Draws the user in
o Adds more usefulness to the information
Edward Tufte – “5 Grand Principles Of Data Visualization”
4. Integrate All Visual Elements
o Use text, images and numbers where appropriate
o Don’t push context to a legend or title o Don’t make the user “learn your system”
5. Content-Driven Design o Quality data
o Relevance o Integrity
“Good information design will never save poor content!”
http://www.jonkolko.com/projectFiles/scad/IACT370_05_TuftePrinciples.pdf
7 Data Driven Summit 2014
1. Display neither more nor less than what is relevant to your message
2. Don’t include visual differences in a graph that do not correspond to realistic
comparison
3. Use the size and location of objects to encode quantitative values
4. Differences in values should be portrayed accurately (start from “0”)
Stephen Few – “7 Core Design Principles For Displaying Quantitative Information”
5. Do not connect values that are discrete,
suggesting a relationship that does not exist
6. Emphasize the information that is most important to your message
7. Augment people’s short-term memory by combining multiple facts into a single visual pattern
“Good data visualization takes the burden of effort off the brain and puts it on the eyes”
http://www.perceptualedge.com/articles/Whitepapers/Visual_Communication.pdf
8 Data Driven Summit 2014
• What is “data visualization”?
• What is “good”?
• Ed Tufte principles
• Stephen Few principles
• Visualizing data
• Chart basics
• Visualization Examples
• BIRT Style Techniques
Agenda
8 Actuate
Corporation ©
9 Data Driven Summit 2014
Chart Basics: “how do I best depict the data?”
Choose the appropriate visual for the information
• Tracking values over time (eg: daily sales for the past quarter):
• LINE CHARTS or AREA CHARTS with the time dimension on the X-axis
• Comparing summarized amounts across categories (eg: transactions by merchant):
• COLUMN or BAR CHARTS
• Comparing a percentage value against the whole (eg: % breakdown of total expense by division):
• PIE or DONUT CHARTS
• Displaying current performance by region/state/country/territory
• MAPS (color coded, sub-graphics overlay)
• Animated Visuals – interaction increases value
10 Data Driven Summit 2014
Chart Basics: “how do I best depict the data?”
Chart Tips
• When category values are too long to display in the X- axis of a column chart…
• …pivot the X-axis to vertical alignment (bar chart)
• On column or bar charts, show bars sorted by value ONLY when you want to convey ranking
• Otherwise X-axis categories will shift when displayed with different contexts
• On a line chart, show data point markers ONLY if you want to convey specific values
• Otherwise it detracts from trend depiction of a smoother
line
11 Data Driven Summit 2014
• What is “data visualization”?
• What is “good”?
• Ed Tufte principles
• Stephen Few principles
• Visualizing data
• Chart basics
• Visualization Examples
• BIRT Style Techniques
Agenda
12 Data Driven Summit 2014
Geospatial
US Unemployment Level
13 Data Driven Summit 2014
Charts on Top
14 Data Driven Summit 2014
Interaction
15 Data Driven Summit 2014
Icons
16 Data Driven Summit 2014
Information Rich
17 Data Driven Summit 2014
Repetition
18 Data Driven Summit 2014
Drilldown
19 Data Driven Summit 2014
• What is “data visualization”?
• What is “good”?
• Ed Tufte principles
• Stephen Few principles
• Visualizing data
• Chart basics
• Visualization Examples
• BIRT Style Techniques
Agenda
20 Data Driven Summit 2014
BIRT Styling Techniques
Visual Styling Enhances Aesthetic Quality and Understanding 3 layers of style control in BIRT:
• Styles
• Granular aesthetic settings
• font-family, font-size, font-color, etc…
• Themes
• Collections of Styles Following a Specific Scheme
• Libraries
• Encapsulation of BIRT components, including Themes
• Centralized Control of Application Look-n-Feel
21 Data Driven Summit 2014
Styles
• Based on Cascading Style Sheets (CSS)
• Options:
• Creating custom styles in your report design
• Predefined styles in your theme
• Custom styles in your theme
• Importing a CSS file into your theme
• Linking to an external CSS file in
your theme
22 Data Driven Summit 2014
23 Data Driven Summit 2014
24 Data Driven Summit 2014
Themes
• A theme is a set of styles applied to BIRT visualizations
• Themes can be defined at
• Report level
• Object level (table, crosstab, chart, label …)
• Can use default themes
• Can create custom themes
25 Data Driven Summit 2014
26 Data Driven Summit 2014
27 Data Driven Summit 2014
Libraries
• Libraries are collections of reusable components
• Data items (connections, data sources, parameters)
• Report Items
• Reusable report items
• Master pages
• Themes
• Aesthetics
• Provides centralized control for Themes and Styles
• Enables Rapid Changes
• Change Styles/Themes in Library Will Update Entire Application
28 Data Driven Summit 2014
Demo BIRT Style Techniques
• CSS
• Library
• Themes
• Styles
29
Actuate Corporation © 2014
30
Actuate Corporation © 2014
31
Actuate Corporation © 2014
32 Data Driven Summit 2014
33 Data Driven Summit 2014