When Augmented Reality meets Big DataCarlos BERMEJO, Zhanpeng HUANG, Tristan BRAUD, and Pan HUI

System and Media Laboratory,Department of Computer Science and Engineering,

The Hong Kong University of Science and Technology,Clear Water Bay, Kowloon, Hong Kong

Email: cbf@cse.ust.hk, soaroc@cse.ust.hk, braudt@ust.hk, panhui@cse.ust.hk

Abstract—We live in an era where we are overloaded withdata, and this can be the key for gaining rich insights aboutour world. Augmented reality (AR) enables us the possibility tovisualize and analyse the growing torrent of data in an interactivecanvas. We can display complex data structures in simpler andmore understandable ways that was not possible before. Big Datais a new paradigm results from the myriad data sources such astransactions, Internet, social networks, health care devices andsensor networks. AR and big data have a logical maturity thatinevitably will converge. The tread of harnessing AR and big datato breed new interesting applications is starting to have a tangiblepresence. In this paper, we explore the potential to capturevalue from the marriage between AR and big data technologies,following with several challenges that must be addressed to fullyrealize this potential.

I. INTRODUCTION

The development of Internet, social network, mobile deviceshave had an impact in the amount data transmitted and sharedacross the globe. Besides, we are facing a new Business erawhere data-driven decisions are better decisions.

Human intuition tells us it will be easier to make sense ofand interact with information if it is merged with the physicalworld [15]. As a modality to display information by overlayingvirtual content on the current view of the world around us,AR enhances the way we acquire, understand, and displayinformation without distraction from the physical world [10].Over the past few years, AR has been progressing by leapsand bounds in terms of technology and its applications.However, even though technologies such as sensing, track-ing, and displaying have improved, AR application patternsbroadly follow the lines of prototypes and demonstrations,such as virtual pop-up objects on 2D markers, and sampledata visualization. A major impediment to AR adoption is thelack of data sources described by MacIntyreetal [15]. AR isdata-hungry and requires more data than most applications.Apart of the application-specific content that users see andinteract with, AR needs to feed applications knowledge aboutits surroundings and descriptions about how it relates tothe application data. Imperative environmental informationmay include geospatial coordinates and models of nearbybuildings, features and semantic descriptions of objects, andlinkage between physical and virtual content. Previous worksacquired data from the legacy database, but the data maybe incomplete or out-of-date due to sparse sensing and theabsence of persistent maintenance. The first AR prototype was

developed by Sutherland in 1960s [21], but it only started todraw attention in the last two decades. A typical AR includesthree characteristics defined by Azuma [1]: Combines the realand the virtual; Interactive in real time; Registered in 3-D.

Applications such as tourism [19], advertisement1, educa-tion [5], and assembly [7] are looking for a way to supplementthe physical world with virtual content rather than replacingit with VR applications. Recently, AR’s popularity has grownon mobile devices. This subset of AR is known as MobileAR (MAR). Pokemon GO2, for example, is a well-knownMAR application that offers location-based AR mobile gameexperience. The Pokemon GO predecessor Ingress3 generatedalmost 2 million US dollars the first days after its release.

The high penetration of technologies spanning mobility, so-cial networks, and the Internet of Things (IoTs) has catapultedthe world in the era of big data. Touted as a game changer, bigdata has been identified by the US government as a researchfrontier that is accelerating progress across a wide range ofpriorities [9]. AR and big data have been around shaping theirown landscapes in various fields for a few years. However, theintersection of two disruptive technologies has not attractedmuch attention yet. The rich insight of big data and noveldisplay modality of AR is promoting the convergence of ARand big data. AR has great opportunities to bring innovation tobig data in terms of visualization and interaction. In this paper,we explore the potential opportunities from the convergenceof both disruptive technologies alongside several challengingproblems that should be addressed. According to a McKinseyreport [16], big data has the potential to reduce product de-velopment and assembly costs by 50% and increases retailersoperating margins by 60%. It is estimated to create savings of300 billion dollars in healthcare every year in the US alone.Big data analysis has been widely used in healthcare4, energysaving, and financial risk analysis [4].

II. AR-POWERED BIG DATA

Interpretation is a major phase of the big data analysispipeline, which is critical for users to extract actionableknowledge from massive and highly complex datasets. As an

1http://www.businessinsider.com/11-amazing-augmented-reality-ads-2012-1

2http://www.pokemongo.com3https://www.ingress.com4https://www.nsf.gov/pubs/2012/nsf12512/nsf12512.htm

intuitive interpretation, visualization transforms plain and bor-ing numbers into compelling stories to help users understandthe data. In addition, big data analysis requires a human-in-the-loop collaboration at all stages of the analysis pipeline[13]. Powerful interactions help to explore and understand thedata more easily and fully, understanding the data is the keyto success in the current data-driven business ecosystem.

A. Data Visualization

With visualization, we turn the big data into a landscapethat we can explore with our eyes. A visual informationmap is useful when we are drowning in information. Datavisualization has the ability to take the complex abstractsymbols and transform them into simpler visual concepts thatwe can quickly understand. Visualization-based data discoverytools have already delivered greater customer and marketinsights to businesses around the world5. It is estimated byGartner that data visualization tools will promote a 30%compound annual growth rate in 2015 [20]. In the past, wehave employed diverse ways of visualizing data using tabularpresentations, interactive bubble charts, treemaps, heatmaps,3D data landscapes, and other types of graphics. The datais generally displayed on flat mediums such as desktops andmore recent mobile screens, which separates the visualiza-tion from the data source and user context. For instance, avirtual array of gauges to display temperatures would failto fully describe the spatial distribution as the temperaturesare strongly related to real world objects in a physical usercontext. In many cases, it is easier to gain insight from thedata if the visualization is embedded in the physical world. Bigdata visualization can be enhanced if an AR layer is overlaidon real-time streaming data or user context. The AR enablesusers to be immersed in a world where the data is much moreeasily understood if it corresponds explicitly to real content,as shown in Figure 1. In this Scenario, since data is usuallydirectly connected with physical context, display simplificationwhich impairs user’s understanding in traditional display is nolonger required. For instance, to find a book among millionsof others in a library, traditional methods require digital 2D or3D maps and floating bubbles for coarse positioning. Poweredwith AR, users are able to see through walls and shelvesto look for indications, e.g. the highlighting contour of thebook. As an indication is usually located according to auser’s current view, it will be much easier to find the book.Collaborating with AR, big data is especially suitable forin-situ visualization and field diagnosis. For example, withrapid growth in information building information modeling(BIM), AR steps in every phase of a building’s life cycle,from building to maintainance [23]. This enables field workersto view the on-site feasibility design and assist constructionwith virtual simulations and clash detection. It also facilitatesassets management in daily inspection activities based on atorrent of data from in-built sensors [12]. To achieve success

5http://www.intel.eu/content/www/eu/en/big-data/data-insights-peer-research-report.html

Fig. 1: Visualization of a numerical flow field with realbuildings makes the influence of the building on windmovement easily understood. (source: http://emcl.iwr.uni-heidelberg.de/research sv.html)

with big data visualization, we need to rethink how to mixdigital data with physical world and present the informationto users. Apart from application-specific requirements, otherconsiderations should be kept in mind. Floating bubbles arewidely used by many AR applications, however, it seems to bepointless and no improvement on a 2D map [23], especiallywhen the data content can not be seamlessly integrated intothe real world. Content should be merged into the physicalworld in a way that users perceive as a real counterpart,just as the 1st and Ten system used by ESPN in Americanprofessional football broadcast. To achieve this, visualizationrequires immense graphic effort and additional informationfor visual occlusion (e.g. something hidden behind a physicalbuilding) and ambient lighting.

B. User Interaction

Today’s world is becoming a canvas for big data from awide range of sources, which profoundly influences people’sperception and understanding of the environment around them.It requires a user-friendly interface to interact with the digitalworld. Traditional user interface design is constrained byfinite physical dimensions. It becomes especially severe whenmobile devices are preferred as a medium to interact withtorrents of data from social media, online transactions, andtelecommunications.

The ability to associate data with the physical world dis-closes the causality between data and reality. In addition,mashing up data from various sources dramatically increasesthe probability of discovering relevant and interesting things.Furthermore, not only the user interface (UI) is a key factor inusers engagement for AR applications, but the user experience(UX), which involves user behaviour and emotions towardsa specific artefact, needs to be considered in any currentapplication design.

A few works [8] further employed AR as user interfaceto redefine and attach new functions to physical objects. It iseven more compelling when AR is integrated with accessoriessuch as AR glasses and AR contact lenses. Wearable andinvisible accessories reduce device intrusion, which providesa hand-free interaction experience. Smart glasses and contact

Fig. 2: Left: Sight, a short futuristic film by Eran May-razand Daniel Lazo to look into a future AR with retinal lenses.Data from sensors, apps, and Internet augment current views(source: http://vimeo.com/46304267); Right: A prototype ofbionic contact lenses with a built-in LED for virtual displaysource7

lenses6, such as Google Glass, made huge strides in this area.With implanted and invisible accessories, AR allows usersto interact with data securely. As the user interface is onlyseen and manipulated by the user alone, it reduces the riskof data and privacy leakage. Data can be viewed throughthe AR interface as personal media or by multiple usersin a collaborative way. As a personal information center, itcollects data from various sources, and displays it on intangibleinterface without physical constraints. In the collaborativemode, multiple users share the same data set and view it fromtheir own angle.

AR applications can enable a new approach to providedata visualization of complex Big Data structures in an easiermanner and provide a better experience for users to interactwith it. Besides, AR can be a very usable canvas to visualizebig data sets as we can see from Sci-fi movies such as Avatar(Figures 2, 3).

III. BIG-DATA-DRIVEN AR

Big data and AR have shaped new business regimes thatwere irrelevant in the past, but the landscape is undergoinga seismic shift with advances in technology convergence andconnectivity. A majority of AR applications are constrainedin an enclosed or pre-compiled environment due to a limiteddataset that is either not available or too sparse to use. Asrapid penetration of mobile devices, social networks, and IoTare generating considerable amounts of data, it makes sensefor big data to enable AR to be more feasible for practicaluse.

A. Retail

AR has been in use in the retail business for the last 5 years.A majority of applications such as Junaio and Wikitude ARbrowsers overlay geospatial-related data on current view tooffer general information. A few AR apps promote virtual ad-vertisements to catch customers’ eyes. However, it is frustrat-ing to boost customers’ interest in products if their behaviorsand preferences are not taken into account. Without adequate

6http://www.engr.washington.edu/facresearch/highlights/ee contactlens.html

Fig. 3: Large data visualization and interaction among multipleusers in the science fiction movie Avatar directed by JamesCameron, which may be portrayal of future user interface withAR. (source: http://www.avatarmovie.com/index.html)

information from customers, AR is less attractive for practicaluse and more like a gaudy, flashy technology. The mobilitytrend preferred for purchase and social communication hasled to the birth of digital consumers8, which also brings usto the era of product digitalization. Digitally active consumershave changed their mode of transaction, communication, andpurchase decision making. Consumers use mobile devices foronline transactions, and social media for advice on what tobuy and where to shop, leaving trails of their performances,states, and decision. It is obviously valuable to explore thelarge amounts of data to understand consumers’ shoppingpreferences and behavior, which helps to tailor promotions andrecommendations to customers. However, it can be even morepromising when it is combined with AR technology. Harnessedwith big data, AR promotes vertical retail to individual con-sumers. A conscious and activated shopping context have animpact on customers’ mental presentation [2], helping them tobe active, informed and assertive in their shopping decisions.

When the new eye gazing and facial expression technolo-gies, and other physiological measurements eventually gaina tangible presence, it will enable us to better understandcustomers’ focus and emotions to provide more accuraterecommendations and advertisements. A few works have usedeye-tracking glasses to collect customer point of gaze in-formation for shopping behavior analysis. According to aMarks & Spencer report [24], people using mobile shoppingchannels spend eight times as much as people shopping instores. Meeting consumers where they are is the key tofuture consumer engagement [22]. With the advantages of highmobility and consistent virtual content provided by big data,AR breaks physical constrains to enhance the virtual shoppingexperience anywhere and any time.

B. Tourism

In terms of environmental awareness, AR is concerned aboutpresenting contextual information and assisting in daily activ-ities, which is particularly helpful when people are unfamiliarwith the environment around them. By highlighting interesting

8http://www.businessinsider.com/11-amazing-augmented-reality-ads-2012-1

features or bringing history to life, AR provides intuitivemeans to enhance a touring experience. As travel is normallyassociated with geo-spatial exploration, most AR applicationsfor travel guides are based on geo-spatial information. Auser’s position is tracked using GPS and built-in sensors, andit is then used to search and locate multimedia informationfrom data sources such as point of interest (POI) databases,geocoded Tweets, and Flickr.

Trends such as the fast deployment of sensor networks andthe growing use of mobile devices and social networks aregenerating large amounts of both structured and unstructureddata. Aggregating and compiling the redundant fragmenteddata helps us to build a detailed and complete environmentalmodel, which enables AR to understand users’ surroundingsbetter even in an open and unfamiliar environment. Accordingto the Business Insider’s survey9, intelligent recommendationis regarded as the most attractive expectation of tourism. Asthe world is getting smarter with big data, it comes with theability of tracking and measuring tourists’ needs and behaviorsto ensure a responsive and intelligent trip experience. Forinstance, personalized travel guide information is overlaid onthe tourists’ current view to avoid being distracted from touristspots and getting lost. Native language signs are automaticallytranslated into readable words which are overlaid on the orig-inal places. To go a step further, information such as locationsof nearby rest sites and restaurants can be recommendedaccording to tourists’ needs based on walking distance andtime.

C. Health Care

When we are making life and death decisions, immediateaccess to relevant and necessary information is of the utmostimportance. AR importance in healthcare industry is attributedto its ability to instantly in-situ display relevant informationwhen required. AR has been used in the medical field fornearly ten years. We have seen AR’s powerful ability of x-rayvision, which has been used to provide contextual cues fordiagnosing patients and learning tools for medical students byprojecting computerized tomography (CT) scans or medicalimages on a current view10. In one example, images of veinsare overlaid on a nurse’s view of a patient’s hand to helpthe nurse insert the IV in one painless attempt. In another ARapplication developed by German research institute Fraunhofer11, a digital overlay of key blood vessels are displayed onthe iPad when the doctor holds the embedded camera on apatient’s body to avoid accidentally cutting them. Although theearly examples have proven AR’s ability to change the health-care landscape, AR’s great lifesaving potential for healthcareindustry can not be fully presented without big data support.

Without adequate data sets, AR is merely for medicaleducation purposes and a bedside manner test. A decisionis strongly made based on doctors’ hunches and experience

9http://www.businessinsider.com/infographic-how-technology-will-eliminate-travel-frustration-2012-8

10https://www.youtube.com/watch?v=VF2mKavngHE11http://sixrevisions.com/user-interface/the-future-of-user-interfaces/

Fig. 4: Corning’s vision of a future operationroom augmented with remote presence. (source:https://www.youtube.com/watch?v=VF2mKavngHE)

rather that the data itself. According to a survey by Man-hattan Research12, 72% of physicians routinely use computertablets every day. With the substantially increasing usageof tablets among physicians and digital care devices amongpatients, paper prescriptions and manual health records arebeing replaced by electronic health records (EHRs). Patientdata is being digitalized, leading to a flood of digital datafrom which medical decisions that previously were based onguesswork and experience can be made based on data itself(Figure 4). Creating a virtual viewfinder to display a pertinenthealth record enables the doctor to quickly access valuableinformation in the context of patients. In-suit visualization ofhistorical illnesses or tissue damage over patients themselveshelps the doctor understand his or her patients better. In future,AR can even visualize a virtual operating room for doctors atdifferent places to diagnose a patient in a collaborative way.

Working closer with ourselves in daily life, AR can besignificantly important for self-tracking health, as wearabledevices that could sense heart rate, blood oxygen, or evencholesterol level become available to everybody. With eachone of us becoming a walking data generator [17], we wouldbe able to keep track of our own basic health statistics. ARdisplays real-time notifications to allow us to immediatelyunderstand our own health condition, and it may even makesuggestions based on health statistics and diet.

D. Public Services

The government is responsible for providing public services,which both produce and consume large amounts of data. Asthe largest spender in any economy, the government has themost diverse channels for collecting data from public services,which includes transportation, social services, national secu-rity, defense, and environmental stewardship, among others.According to a joint global survey by Bloomberg Business-week and SAP [18], 81% of the questioned believe that publicsector will inevitably be transformed by big data. The govern-ment can use big data to provide better services to citizens.Public services, such as transportation and security protection,

12http://www.prnewswire.com/news-releases/new-study-reveals-that-physicians-embrace-patient-self-tracking-202522041.html

will be more efficient and productive if they are delivered di-rectly to individuals in their own contexts with AR technology.For instance, the upcoming traffic flow displayed on a screenwill help drivers avoid traffic accident. Personal informationoverlaid on passengers will enable security specialists tovery quickly verify identification and reduce screening traffic.Augmented government [3] has been proposed to improvegovernment services with AR technology in a few US publicsectors. The AR strategy for government service deliverywill be more promising if it is supported by big data fromsurveillance systems, mobile and sensor networks, and socialmedia. As sensors and wireless networks continue coveringvehicles, it is much easier to collect massive traffic statisticsto make us aware of the traffic situation. For example, carspresent within a range of distance can share GPS positions,speed, and direction information with a vehicular Ad-hocnetwork (VANET). AR can display the information in frontof drivers for performing thread assessment and predictingany potential car crashes. In particular, by harnessing the x-ray vision capability, drivers can see through buildings orvehicles to watch for vehicles positioned in their blind spots.Overlaying essential information such as driver’s license andvehicle’s location and speed directly over the vehicle will alsohelp traffic police to quickly determine whether the driverviolated any traffic rules. A similar strategy can be employedby security agencies to rapidly identify suspects. The city isbeing digitalized by ubiquitous sensor and social networks,which makes it transparent for city managers to look into.To give an example, in a civil engineering maintenance workscenario, a virtual image of a subsurface infrastructure canbe superimposed on a field workers’ vision of the site toenable them rapid perception of the underground networklayout. Field workers can collaborate from different aspects bygiving contextualized views to each supporting role. Individualview is personalized and annotated for each worker’s context,such as electrical-line view for the electrician and plumbing-line view for the plumber, which harnesses the collectiveintelligence of all roles to improve efficiency. In anotherexample, a virtual bird’s eye view directly overlaid on anemergency staff’s vision will greatly assist in the search andrescue of persons trapped in a burning or collapsed building.As civil infrastructure such as electrical and water supplynetworks are getting smarter with IoT, torrents of data canbe used for in-suit visual analysis without field damage byusing AR technology.

IV. CHALLENGES

We have seen an emerging shift in mindsets of big data andAR from industry and business, but there are still significantbarriers to overcome. Several barriers, such as intrusive dis-play, battery life, and highly fragmented data, are practical.Some barriers are conceptual. We should first address thesechallenges before we can see their full potential in action.

Converging big data and AR brings practical technical chal-lenges from both sides. Heterogeneity, scale, and complexityare problems with big data that impede the process of all

phases from data acquisition to aggregation and analysis.AR applications are also constrained to extensive calibration,incomplete reference modelling, and poor environmental sens-ing. A comprehensive discussion of the technical problemsfrom each side is out of the scope of this paper. Interestedreaders can instead refer to Huang’s [10] and Labrindis’ [13]papers for details. Herein we explore emerging practical tech-nological problems and common conceptual barriers inducedby converging both technologies.

A. Timelines

AR applications generally require real-time performance toguarantee fluent user interaction, which requires immediateanalysis results (20 to 7 ms) from Valve software study13.However, large-scale data analysis usually takes longer timesdue to the voluminous and highly fragmented data. Theproblem becomes considerably more challenging when thetrend of minimization in AR devices conflicts with the growingvolume and fragmentation of big data. The cloud is ableto store a large amount of data and handle computationallyintensive tasks within a fixed time cap. A few applications [11]have proven cloud’s ability to meet requirements in a timelyfashion. In addition, offloading computation and data storageenables client-side AR devices to be small and sustainableenough without intrusion.

B. Interpretation

By integrating big data with AR, we acquire data analysisability. However, the ability to analyze data is of limited valueif the results can not be understood by AR. Users of bigdata analytical systems are data scientists while AR users arecustomers without much technical background. The presentdata analysis pipeline has rather been designed explicitly tohave a human in the loop because many patterns are obviousfor humans but difficult for computer to understand [13], how-ever AR prefers to be intuitive and used without interruption.Analysis results require interpretation in the context of AR. Forinstance, the output of a customer behavior analysis system isnormally customer stats, but AR is responsible for how touse the stats. AR should be able to interpret the results aspreferential information so as to provide a recommendationto a customer’s specific context. Although big data is good atdiscovering correlations, especially subtle corrections that arenot possible from a small data set, it does not tell us whichcorrelations are meaningful, while AR requires semanticallymeaningful information to relate to the users’ context. Thereis no easy way for big data and AR to intelligently interpretfor each other. However, a collaborative effort to embraceAR content and provide native APIs for AR to interpretsemantically-tagged data from all data generators is a possiblesolution. A standard data format such as Augmented RealityMarkeup Language (ARML) [15] is an essential step in theright direction.

13http://blogs.valvesoftware.com/abrash/ latency-the-sine-qua-non-of-ar-and-vr/

C. Privacy

There is a growing concern about privacy in the context ofboth AR and big data. In order to provide a personalized rec-ommendation, AR generally requires to access users’ personalinformation and record their location. Studies show that users’identities and their movement patterns have a close correlation[6]. Even though people pay attention to personal information,an attacker can infer private information from their locationinformation [13]. Although the data is fragmented and incom-plete from individual sources, it is interrelated and includesfrequent patterns and redundant knowledge. When data isaggregated from numerous data sources, hidden relationshipsand models can be extracted by data analysis. Privacy isboth a technological and sociological problem. Forceful lawsand regulations may be required to avoid inappropriate useof personal data and malicious AR applications. From thetechnical standpoint, differential privacy is a possible wayof accessing data with a limited privacy risk, however theinformation is reduced too far to be useful in practice. Notonly that, it is ill-suited for dynamically changing data.

V. CONCLUSION

In this paper rather than just an up-to-date survey of big dataand AR technologies, we have broadened our outlook to mergethem to breed new applications. The factors for promotingthe convergence of the two technologies also created somechallenges not previously experienced by either technology.The key principle is to combine the main features of eachtechnology as we devise novel applications using both. Al-though the technologies are still in their infancy, attention hasalways been a necessary component of the convergent strategyof AR and big data. While the concepts are not unfamiliar tous, most participants have not yet harnessed their full potential.Only when we take full action can we get a real picture ofwhat it will take for big data and AR to move from being justhype to becoming real game changers.

ACKNOWLEDGMENT

This research has been supported, in part, by GeneralResearch Fund 26211515 from the Research Grants Councilof Hong Kong and the Innovation and Technology FundITS/369/14FP from the Hong Kong Innovation and Technol-ogy Commission.

REFERENCES

[1] Ronald T Azuma. A survey of augmented reality. Presence: Teleoper-ators and virtual environments, 6(4):355–385, 1997.

[2] Benedict GC Dellaert, Theo A Arentze, and Harry JP Timmermans.Shopping context and consumers mental representation of complexshopping trip decision problems. Journal of Retailing, 84(2):219–232,2008.

[3] C Doolin, A Holden, and V Zinsou. Augmented government: Transform-ing government service through augmented reality. Deloitte ConsultingLLP. URL: http://www. deloitte. com/assets/Dcom-UnitedStates/Local%20Assets/Documents/Federal/us fed augmented government 0606, 13,2013.

[4] Mark Flood, HV Jagadish, Albert Kyle, Frank Olken, and LouiqaRaschid. Using data for systemic financial risk management. In CIDR,pages 144–147, 2011.

[5] Rubina Freitas and Pedro Campos. Smart: a system of augmented realityfor teaching 2 nd grade students. In Proceedings of the 22nd British HCIGroup Annual Conference on People and Computers: Culture, Creativity,Interaction-Volume 2, pages 27–30. British Computer Society, 2008.

[6] Marta C Gonzalez, Cesar A Hidalgo, and Albert-Laszlo Barabasi. Un-derstanding individual human mobility patterns. Nature, 453(7196):779–782, 2008.

[7] Steven J Henderson and Steven K Feiner. Augmented reality in thepsychomotor phase of a procedural task. In Mixed and AugmentedReality (ISMAR), 2011 10th IEEE International Symposium on, pages191–200. IEEE, 2011.

[8] Valentin Heun, Shunichi Kasahara, and Pattie Maes. Smarter objects:using ar technology to program physical objects and their interactions.In CHI’13 Extended Abstracts on Human Factors in Computing Systems,pages 961–966. ACM, 2013.

[9] John P Holdren, Eric Lander, and H Varmus. Report to the presidentand congress: Designing a digital future: federally funded research anddevelopment in networking and information technology. Executive Officeof the President and Presidents Council of Advisors on Science andTechnology, 2010.

[10] Zhanpeng Huang, Pan Hui, Christoph Peylo, and Dimitris Chatzopoulos.Mobile augmented reality survey: a bottom-up approach. arXiv preprintarXiv:1309.4413, 2013.

[11] Zhanpeng Huang, Weikai Li, Pan Hui, and Christoph Peylo. Cloudridar:A cloud-based architecture for mobile augmented reality. In Proceed-ings of the 2014 workshop on Mobile augmented reality and robotictechnology-based systems, pages 29–34. ACM, 2014.

[12] Javier Irizarry, Masoud Gheisari, Graceline Williams, and Bruce NWalker. Infospot: A mobile augmented reality method for accessingbuilding information through a situation awareness approach. Automa-tion in Construction, 33:11–23, 2013.

[13] Alexandros Labrinidis and Hosagrahar V Jagadish. Challenges andopportunities with big data. Proceedings of the VLDB Endowment,5(12):2032–2033, 2012.

[14] Doug Laney. 3d data management: Controlling data volume, velocityand variety. META Group Research Note, 6:70, 2001.

[15] Blair MacIntyre, Hafez Rouzati, and Martin Lechner. Walled gardens:Apps and data as barriers to augmenting reality. IEEE computer graphicsand applications, 33(3):77–81, 2013.

[16] James Manyika, Michael Chui, Brad Brown, Jacques Bughin, RichardDobbs, Charles Roxburgh, and Angela H Byers. Big data: The nextfrontier for innovation, competition, and productivity. 2011.

[17] Andrew McAfee, Erik Brynjolfsson, Thomas H Davenport, DJ Patil, andDominic Barton. Big data. The management revolution. Harvard BusRev, 90(10):61–67, 2012.

[18] Joe Mullich. Closing the big data gap in public sector. SURVEYREPORT— Real-Time Enterprise,(Sep. 2013), 2013.

[19] Dieter Schmalstieg and Daniel Wagner. Experiences with handheldaugmented reality. In Mixed and Augmented Reality, 2007. ISMAR 2007.6th IEEE and ACM International Symposium on, pages 3–18. IEEE,2007.

[20] Dan Sommer, Rita L Sallam, and James Richardson. Emerging technol-ogy analysis: Visualization-based data discovery tools. Gartner, June,17:3, 2011.

[21] Ivan E Sutherland. A head-mounted three dimensional display. InProceedings of the December 9-11, 1968, fall joint computer conference,part I, pages 757–764. ACM, 1968.

[22] P Udhas, S Mittal, and N Chandrasekaran. Six converging technologytrends: Driving a tectonic shift in the business-consumer ecosystem.KPMG, India, 2013.

[23] David Wilcox, Marvin Johnson, and Peter Carrato. Augmented reality:Bringing bim to life. Journal of Building Information Modeling no. Fall,2012:18–19, 2012.

[24] Z Wood. Marks and spencer gambles on bringing inter-net age to the shop floor. The Guardian Online. Retrievedfrom: www. theguardian. com/business/2012/sep/02/marks-andspencer-multichannel-shopping, 2012.