My talk in the technical meeting "Global Burden of Diseases and Scientific Computation in Health". 25-26 September 2015. FIOCRUZ, Rio de Janeiro, Brazil

1. Data Visualization
Workshop on Global Burden of Disease Big Data and Scientific Computation
FIOCRUZ, Rio de Janeiro, Brazil, August 25-27, 2015
Ramon Martinez
Technical Adviser in Health Metrics
martiner@paho.org @HlthAnalysis

2. Contents
• Data Visualization
• Data Visualization Tools
• Data Analytics and Visualization Platforms
o IHME Visualization Platform
o PAHO Visualization Platform
1

3. Why data visualization?
• There is an amazing connection between
human visual sensors and the brain
• Humans have an extraordinary visual
capacity to detect patterns
2

4. Visual perception
• How many nines (9) are in this matrix?
3

5. Visual perception
• How many nines (9) are in this matrix?
4

6. Visual perception
• Can you see any pattern in this table?
5

7. Visual perception
• Can you see any pattern in this graph?
Title of the Presentation6

8. Statistics and Graphs
• Four distinct datasets
• Each with statistics
properties that are
essentially identical
Title of the Presentation7
Mean of x = 9.0
Mean of y = 7.5
Variance of x = 11
Variance of y = 4.13
Nearly identical
correlation and
regression line
Ascombe’s Quartet

9. Statistics and Graphs
8
• Four distinct datasets
• Each with statistics
properties that are
essentially identical
• But when plotted,
they suddenly appear
very different
• Ascombe suggested
the combined use of
graphs and statistics
methods in data
analysis
Ascombe’s Quartet

10. Why data visualization?
• People make better decisions when they’re based
on understanding.
• For information to be understood, it must often be
presented in visual form because patterns, trends,
and outliers require a picture for the human brain to
see and comprehend.
• Data visualization is essential for:
o Data exploration and understanding
o communicating data
o making better decisions
9

11. What is data visualization?
• Data visualization is the graphical display of
abstract information for sense-making or data
analysis, and communication in a way that leads to
understanding for action.
• Its goal is to visualize data in a way that leads to
understanding.
[Stephen Few, 2014]
• Data visualization is a fundamental product from
the Visual analytics process
10

12. What is visual analytics?
• Visual analytics is the science of analytical
reasoning supported by interactive visual interfaces.
• Visual Analytics methods allow people to combine:
11
human flexibility, creativity,
and background knowledge
today’s IT capacities for
data storage, processing,
and visualiziattion
+
to gain insight and make well-informed
decisions in complex problems
=

13. The analytic process
1. The research question – public health situation
2. Define the analytic framework
3. Identification of data sources
4. Create the analytic plan - methods
5. Conduct analysis & interpretation of results
6. Communicate results and findings - storytelling
7. Propose interventions – actions for improving health
12

14. The Cycle of Visual Analytic
13
When people are exploring
and analyzing data, they
enter into an iterative and
non-linear cycle of
1. data acquisition,
2. selection of visual
structures
3. instantiate data views
(graphs, visual forms)
4. analysis and synthesis,
hypothesis building
5. sharing and re-
examination
where at any point they may
need to skip steps, back-up,
seek additional data, or even
start over.
Diagram adapted from the cycle of visual analytic diagram.
Pat Hanrahan, Chris Stolte and Jock MacKinlay. Visual Analysis for Everyone.
Whitepaper. Tableau Software, Inc., 2007
http://www.tableau.com/sites/default/files/whitepapers/visual-analysis-for-
everyone.pdf

15. From Data To Visualization
14

16. Data visualization tools
• There is a constellation of tools for visualizing data
o DATAVISUALIZATION.CH selected tools: http://selection.datavisualization.ch/
o Andy Kirk’s visualization resources: http://www.visualisingdata.com/resources/
15

17. Data visualization tools
• Libraries
o Libraries for adding charts and visualization capabilities to applications
o Required programming or coding. Suitable for software developers
o Use Programming Language directives to connect to data sources
• Web-based tools
o Visualization tools on the web
o Easy to use, no programing required
o Datasets should be upload. Suitable for small-medium datasets
• Data Analytics and Visualization Platform
o is an enterprise-class (complete) solution that provides the tools and methods for:
1. connect to most of the DBMS standards (SQL, Web services, local files, etc.)
2. to data exploration and visual analytics,
3. creation and design of data visualizations and dashboards, and
4. capability for team collaboration and content sharing, and dissemination
o None or minimum programming required
o Self-service approach, empower business users.
16

18. Visual Analytics & Visualization Platform:
high level system architecture
Title of the Presentation17
Source systems /
Data sources
Data preparation
and integration
Data storage &
managements
Data discovery &
visual analytics
Web-based apps /
services for data
dissemination
Data cleansing, transformation, processing, preparation
and integration processes (workflows & scheduling)
DBMS: Data repository with data (internal and external)
ready for analysis and visualizations
Visual discovery and analytic tools and methods. Data
computation. Data visualization sharing and collaboration
Mortality Health Survey
Diagram adapted from BI/DW technical system architecture
Hospital records PH surveillance Population
GBD
Visualizations
Open Data
Portal
Research
Collaboration
site
MoH Web site

19. Technology Components
• Three integral technology components to
achieve analytic and visualization
o Agile data warehouse solution
o Data preparation and integration software
o Visualization and advanced visual analytics
solution
o Content Management System
Title of the Presentation18

20. IHME GBD Visualization Platform
• IHME implemented in-house custom-software development
to create Data Visualizations
• Development team + collaboration w/ researchers and data
analysts
o 1 Product Owner, Data Visualization
o 1 Data Visualization Scientist
o 3 JavaScript Developers
• Development tools:
o JavaScript - programming language
o jQuery - development framework for JavaScript web-based applications
o D3.js - graphics library
o HTML & CSS - hypertext markup language and style sheet for development and design
of web pages
Title of the Presentation19

21. IHME GBD Visualization Platform
Title of the Presentation20 http://www.healthdata.org/results/data-visualizations

22. PAHO Visualization Platform
• PAHO selected off-the-shelf commercial solutions for
analytics and data visualizations
• Data analysts and health professionals from technical areas
are empower to do self-service analysis and produce data
visualizations
• Visual Analytics and Visualization Platform:
o MS SQL Server - database/data warehouse
o Alteryx - data integration and advance analytics software
o Tableau - visual analytics and visualization software
o Joomla - content management system (CMS) / PAHO Web site
o MS SharePoint - content management system for Intranet and collaboration sites
21

23. PAHO Visualization Platform
Title of the Presentation22