The Internet is evolving from a network of comput- ers to a network of devices, e.g., phones, smart meters, traffic cameras, and air quality sensors. In this Internet of Things, large amounts of data generated by everyday objects can often be organized into data streams, where each data stream is a time series of sensor values sampled together. Visualization is an easy-to-use, efficient, and effective method to present this heterogeneous data to large and diverse audiences, and enable its analysis by users without programming background. Although general data-storage and sharing systems for the Internet of Things, like Pachube and Sensor.Network, offer some basic visualizations, they do little to help users understand relations and patterns hidden in the data, nor do they support live updates to the underlying data streams. Other systems, like Biketastic and the Copenhagen Wheel, feature more complex visualizations but are tailored for a specific application domain and do not address heterogeneous data streams. In this paper, we present SNViz, a Web browser-based AJAX application built using Protovis for visual analysis of large, heterogeneous, and live data streams in the Internet of Things. Besides offering panning and zooming for a detailed look at smaller data subsets, SNViz offers brush-and- linking across multiple visualizations. The latter is invaluable in helping users understand and analyze relationships and patterns hidden in the data. Although SNViz currently works by accessing JSON representations of data streams from Sensor.Network over HTTP, it can be extended to work with other data sources (e.g., wireless sensor network devices or smartphones) and even customized for specific applications.
This work was presented at the Internet of Things conference (www.iot2010.org).
SNViz: Analysis-oriented Visualization for the Internet of Things
1. SNViz
Analysis-oriented Visualization for the Internet of Things
Giacomo Ghidini1 2
Dr. Vipul Gupta2 Dr. Sajal K. Das1
1
Center for Research in Wireless Mobility and Networking
University of Texas at Arlington
{giacomo,das}@uta.edu
2
Sun Labs, Oracle
vipul.x.gupta@oracle.com
Urban Internet of Things Workshop 2010
November 29, 2010
IBM Japan Hakozaki HQ/Royal Park Hotel
Tokyo, Japan
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 1 / 26
2. Outline
1 Introduction
2 Visualization for the Internet of Things
3 SNViz
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 2 / 26
3. Introduction
Outline
1 Introduction
2 Visualization for the Internet of Things
3 SNViz
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 3 / 26
4. Introduction
Century of Cities
The vision. . . and some facts
“The 19th century was a century of empires, the 20th century was a
century of nation states. The 21st century will be a century of cities”
– Wellington E. Webb, former mayor of Denver, CO1
Continuous urbanization2
In 2010, 50% of the world’s populations is living in urban areas
This figure will reach 70% by 2050
Most liveable cities3
Mid-sized
Located in developed country with low population density
How can the Internet of Things help provide high quality of life in large cities?
1
Webb, First Transatlantic Summit of Mayors, 2001
2
U.N., World Population Prospects, 2007
3
Economist Intelligence Unit, Global Liveability Report, 2010
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 4 / 26
5. Introduction
Internet of Things (IoT)
The Internet is evolving from network of computers to network of devices4
AT&T added 900,000 connected devices in Q2, totaling 6.7M5
Non-phone consumer products and machine-to-machine (M2M) units
Smart objects
Connection between physical and cyber worlds
Sensors generate large amount of data
The network world is the computer
How to integrate devices, services, and users in the Web?
4
Duffy Marsan, Network World 2010 [3]
5
AT&T News Release, July 23, 2010 [1]
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 5 / 26
6. Introduction
Representational State Transfer7 (ReST)
“The main idea of REST is to design applications which implement
their functionality completely as a set of URI-addressable resources,
with HTTP being the access method for interacting with them. In
such an application, there is no need for any special interface, the
application fully blends into the Web [. . . ]”
– Erik Wilde, UC Berkeley6
6
Wilde, Putting Things to REST, 2007 [9]
7
Fielding, Ph.D. Thesis, 2000 [4]
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 6 / 26
7. Introduction
Sensor.Network8
Data storage and exchange platform
Removes coupling between devices, services, and users
Data stream
Time series of a set of values (e.g., temperature, pressure, acceleration)
Data points can be added to, edited, and removed from a data stream
Users visualize data streams on line charts, scatter charts, maps, . . .
REST-based API
Users access it via a Web browser
Any device can access it via HTTP
8
Gupta et al., IEEE PerCom Workshops, 2010 [7]
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 7 / 26
8. Internet of Things Visualization
Outline
1 Introduction
2 Visualization for the Internet of Things
3 SNViz
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 8 / 26
9. Internet of Things Visualization
Why Visualization?
Scenario
Partially structured data from
heterogeneous sources
Users with limited or no programming
background
Efficiency
Large amount of data presented in
compact form
Effectiveness
Rely on human vision and processing for
analysis
Ease of use
No specialized data analysis software
required
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 9 / 26
10. Internet of Things Visualization
Visualization and the DIKW Pyramid
Data-Information-Knowledge-Wisdom Pyramid9
Information depends on data, knowledge depends on information, . . .
Ascent of DIKW pyramid is required to acquire wisdom from data
Sound visualization can help ascend the DIKW pyramid
9
Ackoff, Journal of Applied System Analysis, 1989
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 10 / 26
11. Internet of Things Visualization
Related Work
Generic frameworks
Sensor.Network10
No interaction among views of data set
Application-specific systems
Biketastic11
Copenhagen Wheel12
Rich visualizations restricted to a specific domain
SNViz tries to bring visualization to the foreground
striking a balance between flexibility and customization
10
Gupta et al., IEEE PerCom Workshops, 2010
11
Reddy et al., CHI 2010
12
MIT SENSEable City Lab, 2010
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 11 / 26
12. SNViz
Outline
1 Introduction
2 Visualization for the Internet of Things
3 SNViz
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 12 / 26
13. SNViz
Problem
Brush-and-linking
Analyze data by
In multiview representation of
brush-and-linking;
multidimensional data, data points
highlighting; and
selected (brush) in one view are
panning and
zooming highlighted (link) across all views
Challenges
Multiple large,
heterogeneous, live
data streams
Multiple values in
multiple charts in a
Web browser
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 13 / 26
14. SNViz
Architecture
Sensor.Network Viz Sensor.Network
Web browser Internet Web server
No standalone app
Scatter chart, RESTful API
line chart, map,... Well-known
protocol Datastream
GET
Chart Chart metadata
HTTP Requests
1:n
1:n
Datastream
GET
Event Bus Event Bus data
HTTP Responses
JSON content
... ...
Datastream ... Datastream
Compact data
Local copy of data representation
No plug-ins required Or any other tech
HTML & Javascript Java Servlets
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 14 / 26
15. SNViz
Visualization Libraries/Toolkits
Graphical systems (e.g., Processing13 )
Low-level abstraction
Tedious for complex work
Visualization systems (e.g., Google Chart Tools14 )
High-level abstraction
Customized solutions not allowed
Does a solution exist between these extremes?
13
Resig, 2010 [8]
14
Google, 2010 [5]
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 15 / 26
16. SNViz
Protovis15
Marks
Graphical primitives (e.g.,
lines, bars, dots)
Data-driven
Mark features function of
data
Extensible
Proper abstraction level
Control re-rendering Figure: Minard’s Napoleon
15
Bostock & Heer, 2009 [2]
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 16 / 26
17. SNViz
Implementation
Client-side AJAX
HTML, JavaScript
JSON wire format
Google Web Toolkit16
Develop Web application in Java
Translate it into HTML and JavaScript
Deploy it to a Web server/Servlet container
Visualizations
Protovis
Google Chart Tools
16
Google, 2010 [6]
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 17 / 26
18. SNViz
Demo 1/2
Bicycle Ride
Sensor.Network Viz Original visualization
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 18 / 26
19. SNViz
Demo 2/2
Tomato Farm
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 19 / 26
20. SNViz
Analysis of SNViz
Flexible visualization for the Urban Internet of Things
Custom visualizations can be easily added
Supports analysis of live data
Panning and zooming
Brush-and-linking
Responsive and compact
Client-side JavaScript processing
855 KB HTML and JavaScript, incl. 479.9 KB Protovis
2.95 seconds to load Web page (if caching enabled)
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 20 / 26
21. SNViz
More Applications
Sun SPOT Visualization
Each Sun SPOT is a data Sensor.Network
Multiple data
stream Sensor.Network Viz
sources
Sun SPOT
WoT Data Source
Employed in the Java One Web browser Internet Web server
Hands-on Lab 2010 Datastream
GET
Chart Chart metadata
Web of Things Visualization HTTP Requests
1:n
1:n
Datastream
GET
Event Bus Event Bus
Internetworked devices data
HTTP Responses
JSON content
implement RESTful protocol Datastream ... Datastream
... ...
Everything is a data stream
(RSS for data)
HTML & Javascript
Personal data stream mash-ups
can be created
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 21 / 26
22. Conclusions
Conclusions
Developed a modular and extensible Web-based visualization for the Web
of Things
Addressed challenges due to large, heterogeneous, and live data streams
Enabled compact representation of data and analysis of patterns and
relationships
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 22 / 26
23. Conclusions
Acknowledgments
Roger Meike and all the folks
on the IoTA (Internet of Things Actualized) research group
at Sun Labs, Oracle
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 23 / 26
24. Conclusions
References
AT&T.
AT&T Adds Nearly 900,000 Connected Devices to Network in 2Q; Now Services Nearly 6.7 Million Connected Devices, 2010.
Michael Bostock and Jeffrey Heer.
Protovis: a graphical toolkit for visualization.
IEEE Transactions on Visualization and Computer Graphics, 15(6):1121–8, 2009.
Carolyn Duffy Marsan.
10 Fool-proof Predictions for the Internet in 2020, 2010.
Roy Thomas Fielding.
Architectural Styles and the Design of Network-based Software Architectures.
PhD thesis, University of California Irvine, 2000.
Google.
Google Chart Tools / Interactive Charts, 2010.
Google.
Google Web Toolkit, 2010.
Vipul Gupta, Poornaprajna Udupi, and Arshan Poursohi.
Early Lessons from Building Sensor.Network: an Open Data Exchange for the Web of Things.
In 8th IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom 2010 Workshops), pages
738–744, March 2010.
John Resig.
Processing.js, 2010.
Erik Wilde.
Putting Things to REST, 2007.
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 24 / 26
25. Conclusions
Credits
Roger Meike’s Blog, http://blogs.sun.com/roger/entry/the_tomato_forest_connecting_external
Protovis Brush + Link, http://vis.stanford.edu/protovis/ex/brush.html
Protovis Minard’s Napoleon, http://vis.stanford.edu/protovis/ex/napoleon.html
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 25 / 26
26. Conclusions
SNViz
Analysis-oriented Visualization for the Internet of Things
Giacomo Ghidini1 2
Dr. Vipul Gupta2 Dr. Sajal K. Das1
1
Center for Research in Wireless Mobility and Networking
University of Texas at Arlington
{giacomo,das}@uta.edu
2
Sun Labs, Oracle
vipul.x.gupta@oracle.com
Urban Internet of Things Workshop 2010
November 29, 2010
IBM Japan Hakozaki HQ/Royal Park Hotel
Tokyo, Japan
Ghidini, Gupta & Das (CReWMaN and Sun Labs) SNViz Urban IoT Workshop – 11/29/2010 26 / 26