Contexts of Information Seeking in Self-tracking and the Design of Lifelogging Systems Ying-Hsang Liu Paul Scifleet Lisa M. Given School of Information Studies, Charles Sturt University Wagga Wagga NSW 2678, Australia {yingliu, pscifleet, lgiven}@csu.edu.au designed to support specified searches (see [7] for an overview). However, the limitation of this model has Abstract become evident when ordinary users have been engaged with various search activities. Because these users have The development of mobile technology and different levels of domain knowledge, problem situa- wearable activity monitors, making it possible tions, and information searching skills, one of the main for people to retrieve data about their daily challenges is to “develop alternative interfaces that activities, is presenting aspects of information meet the needs of wide-ranging sets of users, and mod- seeking behaviour not covered well by previous els and mechanisms for optimally mapping interfaces research. The main objective of this paper is to problem situation” ([49], p. 114). Yet a single-shot to consider how the new information seeking approach has not been able to support different kinds contexts evident in the use of self-tracking ex- of search behaviours in a pervasive computing environ- tend current understandings of the way people ment. need, seek, share and use information. This Most recently the development of mobile technology paper reviews current trends in information re- and wearable activity monitors, making it possible for trieval system design, interactive information people to retrieve data about their daily activities, is retrieval, and human information behaviour re- presenting aspects of information seeking behaviour search as the foundation for a discussion about not covered well by previous research. Studies of the the way that new trends in information seek- process of tracking personal data generated by daily ing contexts and human information behaviour activities, also called body-hacking, self-tracking, self- can inform research. monitoring or quantified self (e.g., [3, 18, 42, 51]), sug- gest that the large amount of data that are captured 1 Introduction automatically need to be summarised so that users can The paradigm of information access as being a single- make sense of recorded data for various tasks, such shot search request is based on the assumption that as recall, reflection and sharing. As indicated in [40] users’ information needs are static, and can be well the self-tracking paradigm can be characterised as (1) represented by query terms and supported by a sin- sensing (to collect raw data); (2) learning (to inter- gle search box. In a review of the interactivity is- pret, recognise or model behaviours using various data sues of information retrieval (IR) system design [59], processing techniques); and (3) informing, sharing, per- it’s suggested that different design decisions can be suading (to develop community awareness by leveraging characterised by searcher characteristics, conceptual social media). The sensing and learning aspects of self- frameworks (e.g., IR models) and system evaluation. tracking have posed tremendous technical challenges, For instance, many IR systems developed under the while the social aspects are equally important for our framework of Boolean retrieval models were specifically understanding of this new information environment. Conceptually the field of lifelogging has few limits Copyright c 2014 for the individual papers by the paper’s au- and has been described as encompassing all the personal thors. Copying permitted for private and academic purposes. This volume is published and copyrighted by its editors. information an individual might wish to keep track of, In: U. Kruschwitz, F. Hopfgartner and C. Gurrin (eds.): Pro- retrieve and reuse in their own life, including emails, ceedings of the MindTheGap’14 Workshop, Berlin, Germany, family photographs, audio recordings, travel itineraries 4-March-2014, published at http://ceur-ws.org and so on [62, 75]. Lifelogging then, as advocated by the notion of personal information curation [74], prioritises 2 IR System Design the importance of personal information in the everyday The capturing of search contexts is important for IR life of people and focuses on the long term management system design because users have difficulty articulating of personal information for its ongoing value to its their information needs (e.g., [7, 11, 48]). Research creators. The opportunities for lifelogging are however on user query formulation has focused on the captur- changing substantially under a paradigm that Cisco ing, analysis and modelling of search contexts through Systems has described as the ‘Internet of Everything’ search and transaction logs from various systems, such [23], where physical devices and objects connected to as OPACs, search engines and social media. One of the Internet, to each other, and to people are providing the major research issues is to design IR systems that the opportunity to collect and share real time data from can effectively support users’ query formulation tasks people (response times, heart rates, gesture recognition by inferring the user’s familiarity with search topics and other personal biological information) in a network and search intents (e.g., [4, 72]). The techniques of of people, processes, data and things. relevance feedback [38], real-time interactive query ex- This study distinguishes the new class of lifelogging pansion [73] and query suggestion [1, 33] have been systems that have been designed to allow people to proposed and evaluated primarily in laboratory set- capture various kinds of personal information about tings. their body’s state (usually about performance and con- However, one of the main issues in IR system de- sumption) to improve their daily self-monitoring, make sign is when and how to provide assistance through informed decisions and gain self knowledge (with spe- direct system intervention, as we learn more about the cific goals of data gathering) [3, 18, 51, 43, 61, 67], searcher characteristics, search goals and contexts from from other classes of personal information management various sources of evidence (e.g., [10, 33, 50, 72]). With systems (e.g. personal desktop archiving systems). For the availability of large amounts of user search data, instance, consumer products such as Nike+ fuelband1 , these user models have been able to customise search Fitbit trackers2 , UP by Jawbone3 and Strava4 have results by making inferences of user characteristics and been developed to track daily activities with the spe- search contexts (e.g., [19, 78]). Since search terms are cific goal of improving personal health and performance. quite sparse descriptions of complex information needs Most studies have focused on automatic data sourcing, and it’s difficult to interpret contextual information data integration and storage, and data processing (e.g., from search data, these user models have not been [17, 70, 27, 80]), whereas some studies have explored able to consider the higher level of information-seeking the notion of lifelong user profiles [63] in support of long goals and information-seeking behaviours. Nonetheless, term goals [69] and modelling of user characteristics the highly contextualised personal information environ- [56]. Despite the fact that some studies have recognised ment of self-tracking and the quantified self, together the importance of contextual information in the design with users’ long-term information-seeking goals and of self-tracking systems (e.g., [16, 29, 42, 43, 61]), the tasks, as discussed in lifelong user profiles and infor- relationship between information seeking contexts and mation filtering systems, see e.g., [9, 63]) provide a use of personal health information for the design of rich setting for the design of self-tracking in lifelogging self-tracking systems is still unclear. systems. The main objective of this paper is to consider how 3 Interactive Information Retrieval the new information seeking contexts evident in the use of self-tracking and new lifelogging systems extend Research on user interaction issues is the bridge be- current understandings of the way people need, seek, tween system-oriented and user-oriented approaches of share and use information. This paper reviews current IR. This thread of research has been known as inter- trends in IR system design, interactive information active information retrieval (IIR). Recent research has retrieval (IIR), and human information behaviour (HIB) been concerned with user interaction at both the levels research as the foundation for a discussion about the of system and interface (see e.g.,[15, 31, 55]). way that new trends in information seeking contexts From the perspective of interactive IR, research on and human information behaviour can inform research. user information problems has concentrated on the- oretical understanding of user search behaviours in interacting with IR systems by considering the user’s search goals, tasks, cognitive state, search strategies 1 http://www.nike.com/us/en_us/c/nikeplus-fuelband and performance (e.g., [8, 30, 57]). 2 https://www.fitbit.com/au/comparison/trackers In a series of studies designed to make the user 3 https://jawbone.com/up interactions with the text as central processes of IR, 4 http://www.strava.com it’s proposed that user search behaviours can be char- acterised by information-seeking strategies, and IR (ELIS) research program [60]. Originating from the systems should be designed by incorporating differ- field of sociology, the notion of way of life has been ent kinds of user search behaviours [5, 8, 79]. In or- effectively used to characterise ordinary people’s ev- der to characterise users’ information seeking strate- eryday life information seeking contexts (e.g., [2, 71]). gies and model intermediaries’ search behaviours, a As a result, research in this area has been extended to mixed-method approach has been adopted to study take into account information seeking in the contexts user-intermediary interactions in professional settings. of hobbies and leisure activities as part of everyday This thread of research has identified purposes of utter- life (e.g., [12, 14]). However, the transfer of concepts ances and focus of a dialog using discourse analysis [6], between different subfields of Library and Information and later developed into user models of shift of focus in Science, such as HIB and IR, has been difficult for some interactive IR [54], shift of user intentions [77], succes- time [39] (see also [24] for further discussions). sive search in information seeking episodes [45, 44] and task-based model for Web searches [37]. User models 5 Discussion developed from IIR studies, however, have not been widely applied to IR system design (see e.g., [50, 58] 5.1 Capturing search contexts for further discussions). From IR perspectives, since the search terms (or termed query terms or queries) are indicators of user informa- 4 Human Information Behaviour tion needs, researchers have investigated the sources and search effectiveness of search terms in naturalistic Studies of HIB are concerned with how people need, mediated search settings [66], or evaluated a technique seek, share, and use information in various contexts. of eliciting more robust terms from user information Research has focused on how information seeking con- need descriptions [36]. More recently some research has texts at various levels influence people’s information been devoted to the evaluation of multi-query search behaviour [13, 34, 76]. More specifically, in recent dis- sessions [32, 35, 47, 64] and consideration of cross- cussion of the development of conceptual modelling session search behaviours [46]. Overall, these studies in HIB research [21, 65], one of the major forces un- have moved beyond the paradigm of information ac- derlying theory development was a focus on the mod- cess as being a single-shot search request because they elling of information behaviours and the contributing consider the changes in user search behaviours and information-seeking situations or contexts that trigger the relationship between search strategies and search information-seeking actions, as exemplified by several effectiveness within and across search sessions. models (see e.g., models in information behaviour re- The user models developed by IIR researchers have search, reviewed in [76]). the potential for informing the design of self-tracking This line of research arises from the cognitive view- in lifelogging systems since they specifically consider point of user studies with an aim to understanding user successive information searches. For example, the mi- interactions with IR systems and informing the design cro level analysis of user goals [77] has indicated that of new information services and systems. The cogni- users are engaged with different information seeking tive approaches of information behaviour emphasise strategies which can be characterised by types of in- the individual characteristics, whereas the social ap- teractive intentions (i.e., the micro level of user goals), proaches focus on the meanings and values associated methods of interacting with information and resources with the social aspects of information behaviour [53]. encountered. Similarly, studies of transmuting succes- More recently, drawing from the systems approach, the sive searches [45, 44] have suggested that behavioural ecological approach of human information interaction characteristics of searches (e.g., the number of unique focuses on how the environmental constraints shape pages visited) can differentiate stages of successive the use of information tools, with the ultimate goal of search. facilitating the conditions where humans interact with As mentioned earlier self-tracking takes place in systems [25]. highly contextualised personal information environ- Researchers have intensively studied information ments that are directly related to the activities (e.g., behaviours of scholars and professionals since these sport, exercise and driving) or health (e.g., heart rate groups have rich information activities within their monitoring and calorie counts) of people seeking to work environments. As such, the research literature know more about themselves. Similar to the design has accumulated a relatively large number of HIB stud- of information filtering systems [9] and the notion of ies of scholars and professionals (see [41] for a compre- lifelong user profiles [63], this contextualised informa- hensive review). More recent research, however, has tion environment involves users or groups of users, with paid attention to ordinary people and their everyday long-term information-seeking goals and tasks. One of life partly due to the everyday life information seeking the challenges in the design of lifelogging systems for self-tracking is how to represent regular user interests IR/IIR/HIB research processes to account for the infor- as user profiles, and how to summarise logged data mation seeking contexts of individuals, addressing the so that users can make sense for various tasks and questions we have raised will be an important driver long-term use. in future research. 5.2 Information seeking contexts and informa- References tion access tool use HIB research is concerned with the contexts of work [1] Agosti, M., Cisco, D., Di Nunzio, G., and how information access tools can be designed to Masiero, I., and Melucci, M. i-TEL-u: A better support work practices. A recent context-rich query suggestion tool for integrating heterogeneous study of the use of PubMed database in support of contexts in a digital library. Lect. Notes Comput. problem-solving activity suggests the importance of Sci.6273 (2010), 397–400. connecting the user’s patterns of information seeking [2] Agosto, D. E., and Hughes-Hassell, S. To- and their associated needs for information access tools ward a model of the everyday life information support in specified contexts [52]. Drawing from senior needs of urban teenagers, part 1: Theoretical people’s health information seeking strategies [26], an model. J. Am. Soc. for Inf. Sci. Technol.57 (2006), image-based retrieval interface for drug information has 1394–1403. been designed to meet seniors’ needs in a health infor- mation context. These are the examples of qualitative [3] Andrew, A. H., Eustice, K., and Hickl, A. fieldwork that are designed to characterise the contexts Using location lifelogs to make meaning of food of information seeking (i.e., environmental constraints and physical activity behaviors. In Proceedings in [25] and how they affect the use of information access of the 7th International Conference on Pervasive tools. Computing Technologies for Healthcare (2013), Per- Within the context of health information seeking, re- vasiveHealth ’13, ICST, pp. 408–411. searchers have used physiological data to taylor health information for people with diabetes [22], investigated [4] Baeza-Yates, R., Hurtado, C., and Men- the relationship between health information behaviour doza, M. Improving search engines by query and stages of change in physical activity and exercise clustering. J. Am. Soc. for Inf. Sci. Technol.58 [28], identified physical activity information needs [68] (2007), 1793–1804. for people with multiple sclerosis, and explored the relationship between health information seeking and [5] Belkin, N. Interaction with texts: Information health status [20]. Overall, these studies demonstrate retrieval as information-seeking behavior. In In- the importance of considering information seeking con- formation retrieval ’93. Von der Modellierung zur texts to support the design of self-tracking tools for Anwendung (Konstanz, 1993), Universitätsverlag long-term goals and potentially behavioural change. Konstanz. With respect to self-tracking system design, re- searchers have developed empirically motivated design [6] Belkin, N. J. Discourse analysis of human in- principles for lifelogging primarily to support mem- formation interaction for specification of human- ory [75]. Some open research questions regarding self- computer information interaction. Can. J. Inf. tracking and new lifelogging system design include: Sci.12 (1987), 31–42. 1. What are the characteristics of personal health [7] Belkin, N. J. Some(what) grand challenges for information behaviour based on understandings of information retrieval. Sigir Forum.42 (2008), 47– information behaviour in everyday life? 54. 2. How do we consider the reflective personal infor- [8] Belkin, N. J., Cool, C., Stein, A., and Thiel, mation practice of individuals? U. Cases, scripts, and information-seeking strate- 3. How do we incorporate social approaches to infor- gies: On the design of interactive information re- mation use and sharing in self-tracking and new trieval systems. Expert. Syst. With Appl.9 (1995), lifelogging system design? 379–395. A significant gap in research to date is an under- [9] Belkin, N. J., and Croft, W. B. Information standing of the information needs of individuals who filtering and information retrieval: Two sides of are seeking to track their own personal data. While the same coin? Commun. ACM.35, 12 (1992), this paper has explored the possibilities for enhancing 29–38. [10] Bendersky, M., Metzler, D., and Croft, [22] Enwald, H. P. K., Niemelä, R. M., W. B. Effective query formulation with multiple Keinänen-Kiukaanniemi, S., Leppäluoto, J., information sources. In Proceedings of the WSDM Jämsä, T., Herzig, K.-H., Oinas-Kukkonen, Conference (2012), ACM. H., and Huotari, M.-L. A. Human information behaviour and physiological measurements as a [11] Borlund, P., Schneider, J. W., Lalmas, M., basis to tailor health information: An explorative Tombros, A., Feather, J., Kelly, D., and study in a physical activity intervention among de Vries, A. P., Eds. Proceedings of the Second prediabetic individuals in northern finland. Heal. International Symposium on Information Interac- Inf. Libr. J.29 (2012), 131–140. tion in Context (2008), ACM Press. [23] Evans, D. The Internet of Everything: How more [12] Case, D. O. Serial collecting as leisure, and coin relevant and valuable connections will change the collecting in particular. Libr. Trends.57 (2009), world. Tech. rep., Cisco, 2013. 729–752. [13] Case, D. O. Looking for Information: A survey [24] Fidel, R. Human information behavior and of research on information seeking, needs, and information retrieval: Is collaboration possible? behavior. Emerald, Bingley, UK, 2012. In Human information interaction: An ecologi- cal approach to information behavior [25], ch. 10, [14] Chang, S. J. L. Information research in leisure: pp. 199–224. Implications from an empirical study of backpack- ers. Libr. Trends.57 (2009), 711–728. [25] Fidel, R. Human Information Interaction: An ecological approach to information behavior. The [15] Cool, C., and Belkin, N. J. Interactive in- MIT Press, Cambridge, MA, 2012. formation retrieval: History and background. In Ruthven and Kelly [55], ch. 1, pp. 1–14. [26] Given, L. M., Ruecker, S., Simpson, H., Sadler, E. B., and Ruskin, A. Inclusive in- [16] Curmi, F., Ferrario, M. A., Southern, J., terface design for seniors: Image-browsing for a and Whittle, J. HeartLink: Open broadcast health information context. J. Am. Soc. for Inf. of live biometric data to social networks. In Pro- Sci. Technol.58 (2007), 1610–1617. ceedings of the SIGCHI Conference (2013), ACM, pp. 1749–1758. [27] Hamm, J., Stone, B., Belkin, M., and Den- nis, S. Automatic annotation of daily activity [17] Daniel, F., Casati, F., Silveira, P., Verga, from smartphone-based multisensory streams. In M., and Nalin, M. Beyond health tracking: A Uhler et al. [70], pp. 328–342. personal health and lifestyle platformealth track- ing. Internet Comput. IEEE.15 (2011), 14–22. [28] Hirvonen, N., Huotari, M.-L., Niemelä, R., and Korpelainen, R. Information behavior in [18] Doherty, A., Kelly, P., Kerr, J., Marshall, stages of exercise behavior change. J. Am. Soc. S., Oliver, M., Badland, H., Hamilton, A., for Inf. Sci. Technol.63 (2012), 1804–1819. and Foster, C. Using wearable cameras to cate- gorise type and context of accelerometer-identified [29] Hopfgartner, F., Yang, Y., Zhou, L. M., episodes of physical activity. Int. J. Behav. Nutr. and Gurrin, C. User interaction templates for Phys. Activity.10, 1 (2013), 22. the design of lifelogging systems. In Semantic Models for Adaptive Interactive Systems (2013), [19] Eickhoff, C., Collins-Thompson, K., Ben- T. Hussein, H. Paulheim, S. Lukosch, J. Ziegler, nett, P. N., and Dumais, S. Personalizing and G. Calvary, Eds., Human–Computer Interac- atypical web search sessions. In Proceedings of the tion Series, Springer, pp. 187–204. ACM WSDM Conference (2013), ACM, pp. 285– 294. [30] Ingwersen, P. Cognitive perspectives of informa- tion retrieval interaction: Elements of a cognitive [20] Ek, S., and Heinström, J. Monitoring or avoid- IR theory. J. Documentation.52 (1996), 3–50. ing health information: The relation to inner in- clination and health status. Heal. Inf. Libr. J.28 [31] Ingwersen, P., and Järvelin, K. The Turn: (2011), 200–209. Integration of information seeking and retrieval [21] Ellis, D. The emergence of conceptual modelling in context. Springer, Dordrecht, The Netherlands, in information behaviour research. In Spink and 2005. Heinström [65], ch. 2, pp. 17–35. [32] Järvelin, K., Price, S. L., Delcambre, L. Web. J. Am. Soc. for Inf. Sci. Technol.64 (2013), M. L., and Nielsen, M. L. Discounted cumu- 1259–1283. lated gain based evaluation of multiple-query IR sessions. In Proceedings of the ECIR Conference [45] Lin, S.-J., and Belkin, N. Validation of a (2008), Springer-Verlag. model of information seeking over multiple search sessions. J. Am. Soc. for Inf. Sci. Technol.56 [33] Joachims, T., Granka, L., Bing, P., Hem- (2005), 393–415. brooke, H., Radlinski, F., and Gay, G. Eval- uating the accuracy of implicit feedback from clicks [46] Liu, J., and Belkin, N. J. Personalizing infor- and query reformulations in web search. ACM mation retrieval for multi-session tasks: The roles Transactions On Inf. Syst.25 (2007), 1–27. of task stage and task type. In Proceedings of the ACM SIGIR Conference (2010), ACM, pp. 26–33. [34] Johnson, J. D., and Case, D. O. Health Infor- mation Seeking. Peter Lang, 2012. [47] Liu, Y.-H. Individual difference, query reformu- lation and search performance in ad-hoc search [35] Kanoulas, E., Carterette, B., Clough, task. In Spiliopoulou et al. [64], pp. 191–198. P. D., and Sanderson, M. Evaluating multi- query sessions. In Proceedings of the ACM SIGIR [48] Liu, Y.-H., and Belkin, N. J. Query reformu- Conference (2011), ACM. lation, search performance, and term suggestion devices in question-answering tasks. In Borlund [36] Kelly, D., and Fu, X. Eliciting better informa- et al. [11], pp. 21–26. tion need descriptions from users of information search systems. Inf. Process. Manag.43 (2007), [49] Marchionini, G., and Komlodi, A. Design of 30–46. interfaces for information seeking. Annu. Rev. Inf. Sci. Technol.33 (1998), 89–130. [37] Kim, J. Describing and predicting information- seeking behavior on the Web. J. Am. Soc. for Inf. [50] Markey, K. Twenty-five years of end-user search- Sci. Technol.60 (2009), 679–693. ing, part 2: Future research directions. J. Am. Soc. for Inf. Sci. Technol.58 (2007), 1123–1130. [38] Koenemann, J., and Belkin, N. J. A case for interaction: A study of interactive information [51] Melanie, S. Sensor mania! The Internet of retrieval behavior and effectiveness. In Proceedings Things, wearable computing, objective metrics, of the SIGCHI Conference (1996), ACM. and the quantified self 2.0. J. Sens. Actuator Networks.1 (2012), 217–253. [39] Kuhlthau, C. Towards collaboration between information seeking research and information re- [52] Mirel, B., Tonks, J. S., Song, J., Meng, trieval. Inf. Res.10 (2005). F., Xuan, W. J., and Ameziane, R. Studying Pubmed usages in the field for complex problem [40] Lane, N., Miluzzo, E., Lu, H., Peebles, D., solving: Implications for tool design. J. Am. Soc. Choudhury, T., and Campbell, A. A survey for Inf. Sci. Technol.64 (2013), 874–892. of mobile phone sensing. Commun. Mag. IEEE.48 (2010), 140–150. [53] Pettigrew, K. E., Fidel, R., and Bruce, H. Conceptual frameworks in information behavior. [41] Leckie, G. J., Pettigrew, K. E., and Syl- Annu. Rev. Inf. Sci. Technol.35 (2001), 43–78. vain, C. Modeling the information seeking of professionals. Libr. Q.66 (1996), 161–193. [54] Robins, D. Shifts of focus on various aspects of user information problems during interactive [42] Li, I., Dey, A. K., and Forlizzi, J. Using con- information retrieval. J. Am. Soc. for Inf. Sci.51 text to reveal factors that affect physical activity. (2000), 913–928. ACM Trans Comput. Interact.19 (2012), 1–21. [55] Ruthven, I., and Kelly, D., Eds. Interac- [43] Lin, A., Foster, J., and Scifleet, P., Eds. tive Information Seeking, Behaviour and Retrieval. Consumer Information Systems and Relationship Facet, London, 2011. Management: Design, implementation, and use. IGI Global, Hershey, PA, 2013. [56] Šajgalík, M., Barla, M., and Bieliková, M. Efficient representation of the lifelong web brows- [44] Lin, S., and Xie, I. Behavioral changes in trans- ing user characteristics. In Shlomo Berkovsky and muting multisession successive searches over the Santos [63]. [57] Saracevic, T. Modeling interaction in informa- [69] Tang, L. M., and Kay, J. Lifelong user tion retrieval (IR): A review and proposal. Proc. modeling and meta-cognitive scaffolding: Sup- ASIS Annu. Meet.33 (1996), 3–9. port self monitoring of long term goals. In Shlomo Berkovsky and Santos [63]. [58] Saracevic, T. Users lost: Reflections on the past, future, and limits of information science. SIGIR [70] Uhler, D., Mehta, K., and Wong, J., Eds. Forum.31, 2 (1997), 16–27. Mobile Computing, Applications, and Services, vol. 110 of Lecture Notes of the Institute for Com- [59] Savage-Knepshield, P. A., and Belkin, N. J. puter Sciences, Social Informatics and Telecom- Interaction in information retrieval: Trends over munications Engineering. Springer, 2013. time. J. Am. Soc. for Inf. Sci.50 (1999), 1067– 1082. [71] Westbrook, L. Crisis information concerns: In- [60] Savolainen, R. Everyday life information seek- formation needs of domestic violence survivors. ing: Approaching information seeking in the con- Inf. Process. Manag.45 (2009), 98–114. text of ‘way of life’. Libr. Inf. Sci. Res.17 (1995), [72] White, R. W., Dumais, S. T., and Teevan, J. 259–294. Characterizing the influence of domain expertise [61] Scifleet, P., and Lin, A. The strategic value on web search behavior. In Proceedings of the of consumer information systems: Changing the WSDM Conference (2009), ACM, pp. 132–141. face of business and consumer connections at Nike. [73] White, R. W., and Marchionini, G. Examin- In Lin et al. [43], pp. 110–134. ing the effectiveness of real-time query expansion. [62] Sellen, A. J., and Whittaker, S. Beyond Inf. Process. Manag.43 (2007), 685–704. total capture: A constructive critique of lifelogging. [74] Whittaker, S. Personal information manage- Commun. ACM.53, 5 (2010), 70–77. ment: From information consumption to curation. [63] Shlomo Berkovsky, Eelco Herder, P. L., Annu. Rev. Inf. Sci. Technol.45 (2011), 1–62. and Santos, O. C., Eds. UMAP 2013 Extended Proceedings (LLUM 2013: LifeLong User Mod- [75] Whittaker, S., Kalnikaitė, V., Petrelli, elling) (2013), vol. 997 of CEUR Workshop Pro- D., Sellen, A., Villar, N., Bergman, O., ceedings. Clough, P., and Brockmeier, J. Socio- technical lifelogging: Deriving design principles [64] Spiliopoulou, M., Nürnberger, A., and for a future proof digital past. Human–Computer Schult, R., Eds. Report of the Symposium “Ler- Interact.27, 1-2 (2012), 37–62. nen, Wissen, Adaptivität 2011” of the GI Special Interest Groups KDML, IR and WM (Magdeburg, [76] Wilson, T. D. Models in information behaviour Germany, 2011), Fakultät für Informatik, Otto- research. J. Documentation.55 (1999), 249–270. von-Guericke-Universität Magdeburg. [77] Xie, H. I. Shifts of interactive intentions and [65] Spink, A., and Heinström, J., Eds. New Direc- information-seeking strategies in interactive in- tions in Information Behaviour. Emerald, Bingley, formation retrieval. J. Am. Soc. for Inf. Sci.51 UK, 2011. (2000), 841–857. [66] Spink, A., and Saracevic, T. Interaction in [78] Yom-Tov, E., and Diaz, F. The effect of so- information retrieval: Selection and effectiveness cial and physical detachment on information need. of search terms. J. Am. Soc. for Inf. Sci.48 (1997), ACM Transactions on Inf. Syst.31 (2013), 4–19. 741–761. [79] Yuan, X., and Belkin, N. J. Investigating in- [67] Swan, M. Emerging patient-driven health care formation retrieval support techniques for different models: an examination of health social networks, information-seeking strategies. J. Am. Soc. for consumer personalized medicine and quantified Inf. Sci. Technol.61 (2010), 1543–1563. self-tracking. Int. J. Environ. Res. Public Heal.6 (2009), 492–525. [80] Zhou, L. M., Gurrin, C., and Qiu, Z. Zhiwo: Activity tagging and recognition system for per- [68] Sweet, S. N., Perrier, M.-J., Podzyhun, C., sonal lifelogs. In Proceedings of the ACM ICMR and Latimer-Cheung, A. E. Identifying physi- Conference (2013), ACM, pp. 321–322. cal activity information needs and preferred meth- ods of delivery of people with multiple sclerosis. Disabil. Rehabilitation.35, 4 (2013), 2056–2063.