Visualization Viewpoints Editor: Chasing the Negawatt ... - igva2012

Visualization Viewpoints

Editor: Theresa-Marie Rhyne

Chasing the Negawatt: Visualization for Sustainable Living Lyn Bartram, Johnny Rodgers, and Kevin Muise Simon Fraser University

8

E

nergy and resource management is an important and growing research area at the intersection of conservation, sustainable design, alternative energy production, and social behavior. A significant part of energy management involves the buildings we occupy. Advances in design, materials, and automation have resulted in more efficient buildings, and alternative energy sources have changed the role of buildings and their occupants from simple consumers to partial producers of energy. However, effective energy reduction strategies will require not only efficient buildings but also approaches that help occupants modify their energy use behaviors. Energy consumption can be significantly reduced by simply changing how occupants inhabit and use buildings, with little or no additional costs.1 Reflecting this fact, an emerging measure of grid energy capacity is the negawatt: a unit of power saved by increasing efficiency or reducing consumption. Simply put, the cheapest watt is the one that’s never created. Visualization clearly has an important role in enabling residents to understand and manage their energy use. This role is tied to providing real-time feedback of energy use, which encourages people to conserve energy.2 An array of new technologies could provide detailed, localized, and up-to-the minute data about resource and energy use. Bidirectional monitoring with “smart meters” lets residents track their energy use in detail and in near real time. Pervasive and networking technologies underpin the development of in-home networks of devices and appliances. The Internet and social networks connect occupants with a wider community of people pursuing similar goals around sustainability and enable them to share strategies, incentives, and successes. This proliferation of data has spawned a variety of visual-feedback tools. Although there’s both great potential and great interest in promoting behavioral change with these tools, we still know little about how to design, situate, and integrate them to help people make more efficient energy May/June 2010

use decisions. Current applications often require too much effort and management from occupants and aren’t integrated well with the set of information tools people already use. In addition, their design often doesn’t take into account how they will actually fit into home life.3 The challenge is to understand not only what kinds of visualizations are most effective but also where and how they fit into a larger information system to help residents make informed decisions. We need simple yet effective information systems to help users understand their energy use without having to become technology experts, electrical engineers, or control room operators. Our research seeks to find ways that information visualization can contribute to this goal. Here, we examine the effective display of home energy-use data, using examples from our work on North House, a net-zero solar-powered home that placed fourth in the 2009 Solar Decathlon (www.solardecathlon.org). (“Net zero” means that the house produces at least as much energy as it consumes over the course of a year.) We developed this house in collaboration with the University of Waterloo and Ryerson University and in ongoing collaboration with BC Hydro, British Columbia’s public power utility. In particular, we examine the Adaptive Living Interface System (ALIS), the North House’s information backbone.

Avoiding the Control Room Until recently, energy visualizations tended to fall into one of two categories: ■

■

the highly technical displays familiar to building engineers for tuning building performance or simplified displays of aggregated energy consumption over time usable by nonexperts (and familiar from the monthly electric bill).

These “traditional” data visualizations have been joined by a plethora of new approaches, as we

Published by the IEEE Computer Society

0272-1716/10/$26.00 © 2010 IEEE

Figure 1. The Insight Views energy dashboard. Such applications employ a “resident as interested manager” model, which doesn’t always work for residential energy users. (Source: Fat Spaniel Technologies; used with permission.)

show later in this section. The combination of these tools can result in a confusing array of unconnected devices and information tools that require the resident to access, manage, and integrate a confusing stream of information. In effect, this turns the home into a sort of “control room.” The most ubiquitous energy visualizations are based on the model of the “resident as interested manager.” A common application is the energy dashboard. Examples include Fat Spaniel Technologies’ Insight Views (see Figure 1), EcoGauge’s Building Dashboard (used in the first net-zero home in Boulder, Colorado), and Pulse Energy’s Pulse Energy Management Software. These systems provide sophisticated displays and data access tools that support search, analysis, and some prediction of energy use impacts. They cater to several tasks that are core aspects of an informed analytical understanding of home energy use. Simpler dashboard tools are emerging that let people track their energy use and activity on personal Web pages (for example, www.stepgreen.org and the Google PowerMeter). Energy-monitoring dashboards for mobile phones are also being developed but are currently largely proof-of-concept ideas (notably Kaleidoscope’s Current State). Although well-designed dashboards are excellent for analysis, they suffer some shortcomings. The challenge is that people engage with their homes very differently from how a professional manager

monitors and administers a building.4 Residents often have a less sophisticated level of understanding of their home’s electrical system and have limited time in which to make energy-related decisions. They make those decisions from both a strategic, long-term perspective and a tactical, in-the-moment assessment. Such decision-making involves monitoring, awareness, understanding, sensemaking, and control. Looking up and interpreting a dashboard doesn’t support the at-a-glance sensemaking required to facilitate decision-making in the moment. These activities require support at different levels of detail, at varying degrees of attention, and in contextually appropriate ways (both temporally and spatially). Along with energy consumption and cost, residents are concerned with comfort, aesthetics, personalization, and social expectations. So, the resident-as-interested-manager model holds for some types of visualizations and for supporting some kinds of tasks. However, when developing a cohesive system of visualizations to promote sustainability in the home, we must recognize many factors at play in residents’ activities and decision-making. Products such as the Kill A Watt EZ (see Figure 2) take another approach. They’re dedicated energy-monitoring units attached to a particular appliance or outlet, functioning like the old analog meters. They provide purely numerical output, in both electricity consumed and money spent. To IEEE Computer Graphics and Applications

9


Figure 2. The Kill A Watt EZ. This device, which attaches to an electrical outlet, measures energy use but doesn’t convert that data into meaningful information directly related to its user. (Source: P3 International Corp.; used with permission.)

make sense of the data, residents must map these simple values into meaningful information that corresponds to their overall consumption. Such devices have further issues, such as where to locate them in the home. Because these devices are small units with comparatively low-resolution displays, they require focused attention to be read, yet they must be placed at a wall outlet or power strip. Furthermore, they don’t integrate with the residents’ current information “ecosystem” (although do-ityourself projects such as the Tweet-a-Watt seek to address this shortcoming by enabling the Kill A Watt to broadcast and display its collected data wirelessly via Web services). Alternative techniques include ambient visualizations that aim to communicate information without requiring active attention.5 Examples include incorporating displays of energy use into household items such as clocks or power cords and developing abstract informative art as shared ecovisualizations.6 However, these are isolated approaches that haven’t been incorporated into a holistic information ecosystem. If we change our concept of this visualization audience from “interested manager” to “aware resident,” it becomes apparent that the controlroom metaphor used in engineering displays and 10

May/June 2010

building manager tools might be both overly complex and insufficient. Instead, design decisions in visualizing energy use should support residents in decision-making tasks that integrate with their patterns of everyday life and aesthetic preferences. We envision this approach as being more holistic than the currently available tools, emphasizing development of an information and visualization ecosystem. When considered in this context, the design and implementation of visualizations to support sustainable decision-making becomes more than simply refactoring information into casual information visualization for nonexpert interpretation. It involves questions of appropriate context, location, and interaction. (For example, if a person is trying to understand how much hot water he or she uses in a daily shower, a financial graph on the PC is perhaps less effective than a visual readout on the shower tap). A further related consideration is that the typical energy use models for the visual-analysis interfaces used by building managers and engineers rely on a certain level of knowledge of electrical systems. Examples of this knowledge include ■■ ■■

■■

the difference between power and energy, the units in which each is expressed (watts and watt-hours), and how different appliances and energy systems consume resources.

The typical resident might not fully understand these things, so energy use information must be in terms that suit a variety of mental models. For example, expressing energy use in terms of money is a common way to translate abstract concepts of energy use into easily understood concrete terms. Further models can express environmental impact (that is, pounds of CO2 or acres of forest) or translate energy use into commonly understood contexts (“equivalent to four dishwasher cycles”).

Developing a Visualization Ecosystem Visualization in a home setting involves aspects not normally considered in the creation of traditional screen-based analytical tools. Perceptual issues abound. Viewing distance varies: the same device might need to support local and remote visualization. Lighting conditions vary: digital displays are notoriously washed out by sunlight, but people like naturally lit environments. Aesthetic factors are important: how do these visualizations and information appliances cohere with the home’s design and ambience? Along with these go affective constraints: how do the displays stimulate emotional

responses, and are those responses desirable? Can they be negative, if the goal is to encourage energy conservation? There are also ergonomic issues of access: where do we place these visualizations to effectively communicate with residents during their daily activities? How do the visualizations relate to the regular flows of people and energy use over different time periods? Finally, how long are residents willing to spend analyzing and adjusting their energy use? With this in mind, we can’t consider visualizations only discretely; they must form part of a coherent information ecosystem. One route to take in incorporating visualization into everyday activities is to exploit the tools that residents already use. For example, we can embed feedback mechanisms in or near appliances and can create visualization interfaces for mobile computing devices such as laptops and smart phones. Traditional screen-based interfaces will certainly continue to play a primary role as a hub for visualization and analysis, but we can’t expect residents to refer to their computer every time they make an energy use decision. We borrow from Georgina Wood and Marcus Newborough’s comprehensive framework of factors influencing the design of energy-information displays.7 They consider display location, motivational factors, display units (such as kilowatthours, dollars, or grams of CO2), display methods (numerical and diagrammatic), time scale, and category of use. To this framework we add social interaction, personal milestones, and community involvement in the motivational factors. Studies on comparative feedback of sustainability milestones have shown the value of explicitly representing and encouraging those social interactions.8

The Adaptive Living Interface System We encountered these challenges when designing interactive systems for North House. In this netzero alternative-energy home, the resident is both user and manager of a green dwelling. So, the user requires both specific information to help the building work more efficiently and a way to control the house systems. We combined these functional needs with awareness of social factors, such as the potential for community involvement in sharing and comparing energy conservation strategies. To that end, we developed ALIS, a distributed system of interfaces and visualization tools to help the resident become aware of energy use patterns. ALIS includes data-rich views such as dashboards, monitoring and awareness tools, device controls, and state feedback. It also includes aesthetically

Figure 3. The Adaptive Living Interface System (ALIS) Dashboard provides access to high-level information regarding home energy use.

Figure 4. The Resource Usage interface integrates Pulse Energy software to provide retrospective and predictive views of home energy use. (Source: Pulse Energy; used with permission.)

motivated ambient displays. These are variably implemented on four technology platforms: PCs, mobile smartphones, embedded touch panels, and a light-based informative-art display.

The Traditional Approach In ALIS, a PC application offers the most detailed access to feedback about home resource consumption. It gives the resident access to a high-level house Dashboard (see Figure 3), a detailed analytical Resource Usage interface (see Figure 4), and a Neighborhood Bulletin in which residents can share and compare their energy conservation techniques with others in their local community (see Figure 5). The application integrates visualization components in multiple formats: IEEE Computer Graphics and Applications

11


displays are simplified; an innovative and playful “spinner” interface lets residents select pairs of performance variables through slot-machinestyle interaction to compare and graph energy use data on demand. The mobile application also lets residents keep in touch with their neighborhood network, as we discuss in more detail later. Furthermore, it provides alerts about house status and energy consumption and production thresholds. Letting residents interact with their house data through their mobile device is one way ALIS integrates with users’ existing digital ecosystem.

Embedded Displays

Figure 5. The Neighborhood Bulletin lets residents compare energy use and share conservation tips.

Three touch panels are embedded in the home: one in the kitchen backsplash and two at entryways in other parts of the home. These function as control panels and dynamic feedback displays. A prototype “splash screen” for the kitchen panel shows the house’s net-zero status over the past 24 hours and indicates at a glance how the house is doing (see Figure 7). It’s visible at short and medium viewing distances and in variably lit conditions. Behind the splash screen are the controls interface and a subset of the information visualizations in the PC application. The other two panels provide local control and feedback.

Informative Aesthetics: Ambient Canvas

(a)

(b)

Figure 6. ALIS mobile-application visualizations: (a) a line graph showing energy production and consumption over the past week and (b) a data table indicating net-zero activity. ■■

■■

■■

numerical financial data and graphs in the Dashboard, comparative and scalable performance data in the Resource Usage interface, and simplified comparative performance data in the Neighborhood Bulletin.

Because the application runs in a Web browser, these views are available anywhere from a network connection, allowing both local and remote review and analysis.

Mobile Displays A mobile application (designed for the Apple iPhone or iPod Touch) provides residents with a subset of the Web application tools (see Figure 6). Graphing 12

May/June 2010

The Ambient Canvas (see Figure 8) is an awareness visualization that provides feedback on occupant performance in three primary measures of sustainability: net-zero performance, water consumption, and progress toward a user-accepted challenge (a performance goal). This feedback takes the form of curved bands of light on the kitchen backsplash. The light’s intensity indicates the performance level for a specific measure. For example, if the house has been a net producer of energy over the past 24 hours, that part of the Ambient Canvas will glow, with a more intense glow linked to higher performance. It subtly conveys feedback on performance and energy efficiency without requiring active attention from the occupant, whereas the other ALIS components offer more task-focused interfaces. Similar to calm displays such as the InfoCanvas,5 the Ambient Canvas aims to implicitly communicate information through a unique, aesthetically pleasing installation integrated into the home’s structure. We refined the design through many stages with the North House architects. Their concerns about dissonant appearance gave us a strong grounding in how challenging it is to embed informative art into an existing design aesthetic.

Besides this display, the prototyping process included three other proposed embedded technologies: ■■

■■

■■

“organic wallpaper” that provides feedback on energy use through the activation of thermochromatic ink, sonic sculptures that provide aural cues about energy use thresholds and status, and light-based feedback on bathroom water use that maps light color and intensity to water use thresholds.

Each technology seeks to provide ambient feedback about resource consumption to the home’s residents in forms that are nonintrusive, naturalistic, and beautiful. These represent one end of a feedback spectrum that ranges from ambient to explicit.

Social Interaction: The Neighborhood Network Finally, we’ve integrated social-networking features into the PC and mobile phone applications. The Neighborhood Network encourages competition, comparison, and collaboration between community members. Occupants can see a historical view of their energy consumption compared to a community average and set conservation goals, letting them compete to achieve those goals while sharing tips and comments. Those who follow through with their commitments receive awards (digital trophies) that are displayed to other community

Figure 7. A net-zero visualization “splash screen” for the kitchen panel. This screen shows the house’s status over the past 24 hours.

members. These ecotrophies not only apply to the challenges but also recognize the degree to which the participant engaged with the community by posting conservation advice and replying to other members’ feedback. These features are informed by research in social and environmental psychology. Henk Staats and his colleagues demonstrated that groups that discussed conservation-related issues, made reduction commitments, and received comparative feedback reduced their energy consumption compared to a control group.9 E. Scott Geller notes that providing occupants with rewarding feedback, especially from their peers, increases the likelihood that they will repeat a conservation behavior.10 Such studies serve to inform the design of community-based Figure 8. The Ambient Canvas provides feedback on occupant performance regarding net-zero performance, water consumption, and performance goals. The light’s intensity indicates the performance level.

IEEE Computer Graphics and Applications

13


visualization systems by expanding the concept of information visualization to include meaningful social interactions. By enabling occupants to respond to others and discuss their energy conservation activities, we hope to further encourage ecologically sustainable behaviors.

Understanding Success and Avoiding Failure A critical issue in designing these environments is how to evaluate them. Piecemeal usability tests and visualization experiments on parts of the system can help avoid the most obvious mistakes but are inadequate to examine contextual use. Participatory design workshops, scenario walkthroughs, usability studies, and perceptual tests in naturallight settings helped inform the design of the visualizations and interfaces. However, testing these methods in context requires veridical settings, and we can’t simply construct different houses to evaluate different configurations. We’re exploring two approaches. We plan to use North House’s successor as a “living lab” to evaluate these approaches in context. A more speculative approach involves using immersive technologies such as virtual models to more flexibly model different living environments with ALIS components distributed throughout.

G

reen building methods combined with pervasive visualization technologies could be a powerful vehicle for helping to build a conservation ethic. However, simply transferring current approaches into the residential environment is inappropriate. The challenges are legion, not the least of which is how to evaluate these new approaches. Here we’ve argued that the best way to address these issues is to take a broader systemsdesign perspective, opening the boundaries of what’s traditionally considered visualization. This is a fertile research area. We hope the outcomes will include better information tools and interfaces for making energy use decisions, guidelines on designing and locating particular displays in the home, and validated processes for incorporating occupant energy awareness into new-home design and existing-home retrofits.

References 1. R. Woodbury et al., Buildings and Climate Solutions, Pacific Inst. for Climate Solutions, 2008. 2. J.E. Petersen et al., “Dormitory Residents Reduce Electricity Consumption When Exposed to RealTime Visual Feedback and Incentives,” Int’l J. 14

May/June 2010

Sustainability in Higher Education, vol. 8, no. 1, 2007, pp. 16–33. 3. M. Shipworth, Motivating Home Energy Action: A Handbook of What Works, Australian Greenhouse Office, 2002. 4. M. Chetty, D. Tran, and R.E. Grinter, “Getting to Green: Understanding Resource Consumption in the Home,” Proc 10th Int’l Conf. Ubiquitous Computing (UbiComp 08), ACM Press, 2008, pp. 242–251. 5. T. Miller and J. Stasko, “Artistically Conveying Peripheral Information with the InfoCanvas,” Proc. Working Conf. Advanced Visual Interfaces (AVI 01), ACM Press, 2002, pp. 43–50. 6. T.G. Holmes, “Eco-visualization: Combining Art and Technology to Reduce Energy Consumption,” Proc. 6th ACM SIGCHI Conf. Creativity & Cognition, ACM Press, 2007, pp. 153–162. 7. G. Wood and M. Newborough, “Energy-Use Infor mation Transfer for Intelligent Homes: Enabling Energy Conservation with Central and Local Dis plays,” Energy and Buildings, vol. 39, no. 4, 2007, pp. 495–503. 8. J. Mankoff et al., “Leveraging Social Networks to Motivate Individuals to Reduce Their Ecological Footprints,” Proc. 40th Ann. Hawaii Int’l Conf. System Sciences (HICSS 07), IEEE CS Press, 2007, p. 87. 9. H. Staats, P. Harland, and H.A.M. Wilke, “Effecting Durable Change: A Team Approach to Improve Environmental Behavior in the Household,” Environment and Behavior, vol. 36, no. 3, 2004, pp. 341–367. 10. E.S. Geller, “The Challenge of Increasing Proenviron ment Behavior,” Handbook of Environmental Psychology, R.B. Bechtel and A. Churchman, eds., John Wiley & Sons, 2002, pp. 525–540.

Lyn Bartram is an assistant professor at Simon Fraser University’s School of Interactive Arts & Technology, where she directs the Human-Centered Systems for Sustainable Living research group. Contact her at [email protected]. Johnny Rodgers is a master’s of science student at Simon Fraser University’s School of Interactive Arts & Technology, in the Human-Centered Systems for Sustainable Living research group. Contact him at [email protected]. Kevin Muise is a graduate of the master’s of arts program at Simon Fraser University’s School of Interactive Arts & Technology. Contact him at [email protected]. Contact department editor Theresa-Marie Rhyne at [email protected].