VISUAL ANALYTICS
A GROUP 4 PRESENTATION
REMEMBER?
Bike
Blanket
Detergent
Calculator
Kettle
Dustbin
Photocopier
Speakers
Kabali
Wallet
Chalk
Newspaper
Pencil
Apple
photo-frame
REMEMBER?
REMEMBER?
Which is the most stable company to invest in?
2357
980
1628
654
3546
1728
673
889
1124
1526
987
564
972
2435
2738
ANSWER?
Which is the most stable company to invest in?
A
4000
3000
2000
1000
0
1
INTRODUCTION
Data- Structure, Pattern, Connection
Seeing and Exploring
Visual Analytics - Science of analytical
reasoning facilitated by interactive visual
interfaces
Primary Goal- Build better tools, Find better
analytical techniques to take advantage of
human visual and cognitive problem solving
abilities
FEATURES
A single application for BI, data exploration and
analytics
Innovative visualization techniques
Collaboration and storytelling
Autocharting
Text sentiment analysis
Interactive charts, dashboards and reports
Robust report design
Integration with mapping technologies
Mobile BI
Powerful, easy-to-use analytics
Flexible, Scalable Deployment Options
Process
Types of Visual Representations
Interactive Modelling Speed-of-Thought Analysis, What-if
Analysis and Forecasting, Rapid Scenario Planning
Scorecards Personalized scorecards, Measure against goals,
At a Glance information on Business Performance, Convey
information in intuitive format
Dashboards Support what-if scenarios, Drill down
capabilities, Visualize Key Performance Indicators (KPIs)
Management Reporting KPI monitoring, Briefing Books,
Summarized views
Operations and Transaction Reporting Alerts, summarized
views of day-to-day activities of the batch and raw data
reporting
Major Players
Market Share
Other
12%
Microsoft PowerPivot
6%
Tableau
9%
QlikTech
48%
SAS JMP
8%
Spotfire
17%
Scope
Application
Physics and Astronomy
Application
Business
Application
Software Analytics
Application
Biology, Medicine and Health
Application
Mobile Graphic and Traffic
Visual Analytics for
Healthcare Associated
Infections
Healthcare Associated Infections
(HAI)
Definition: Infections that patients acquire during
surgery or treatment for medical conditions in
hospitals/nursing homes
Example: MRSA, Clostridium difficile, Norovirus etc.
Infections Causes: Bacteria, Fungi, Viruses, Parasites,
Prions, etc.
Problems:
1. HAIs result in extended hospital stays, contribute to
increased medical costs and are a major cause of
morbidity and mortality
2. lead to logistic and economic problems for healthcare
services
Objectives
Monitor infections and detect outbreak by:
1. Analyzing and understanding the infection
2. Visually analyse which ward specialities, wards or
hospitals buildings
a. pose more infection risks
b. have an effective infection control strategy or record
3. Monitor patients status to detect:
a. Anomalie
b. Clusters
c. Patterns
d. Relationships
e. Trends
Dataset
Dataset of patient information of an example
infection:
4 year period
447 patients
8479 admissions
10 hospital buildings
127 wards
14 medical specialities
Patient Admissions
Patients and their associated
admissions.
Each horizontal line represents a
patient and the color
corresponds to the hospital
building into which s/he was
admitted.
The number of admissions for
each patient varies widely.
Some patients were only admitted
once while some were admitted
more frequently, e.g. 755 times.
Patient Hospital Building Wise
Number 0f Patients Admitted in Each Hospital
CH admitted the
greatest number of
patients over the 4
year
Period while CI had
the fewest patients
admitted.
Number of Infections in Each Hospital Building
The highest increase in
positive tests was
registered in hospital
building JR.
This could be due to a
relatively higher numbers
of patients going through
this hospital building or
the particular conditions
treated at this building.
Number of Tests in Each Hospital Building
The total number of tests performed at JR
building greatly out-numbers the number of tests
performed in the other hospital buildings.
Patients Hospital Stay Scatter Plot
Number of days a patient
stays in a ward, each dot
represents a patient (xaxis) and the number of
days s/he stayed (y-axis)
and the color represents
the hospital building in
question Stays vary from
day-visits to 147 days
Number of Days in the Ward before Tested
Positive: Scatter Plot
The number of
days that the
patients stayed in
the ward before
tested positive
varies from 0 to
77 days.
Aerospace Engineering Research - Georgia
Tech.
Objectives: Create mapping between two air traffic
control initiatives (Operational Improvements)
Traditional Way:
1. Using Microsoft Word and a search function
2. Identifying keywords and matching them
one by one
Jigsaw:
1. Modification of Documents
2. Developing entity types
3. Analysis: Graphical view and List view
approach
Latest Developments
Visual Analytics for Financial time-series data
SAS Visual Analytics 7.3
Application of VA in Financial Stability
Modelling
Visual Analytics: A new way to manage data
deluge in E-Discovery
BI and Visual Analytics on Mobile: Slight Growth
Reference
http://www.vismaster.eu/book/
chapter-2-visual-analytics/
Daniel A. Keim, Florian Mansmann,
Jim Thomas, Jorn Schneidenwind,
Hartmut Zeigler, Visual Analytics:
Scope and Challenges
