=Paper= {{Paper |id=None |storemode=property |title=An Activity Awareness Visualization Approach Supporting Context Resumption in Collaboration Environments |pdfUrl=https://ceur-ws.org/Vol-743/ASTC2011_Paper3.pdf |volume=Vol-743 }} ==An Activity Awareness Visualization Approach Supporting Context Resumption in Collaboration Environments== https://ceur-ws.org/Vol-743/ASTC2011_Paper3.pdf

An Activity Awareness Visualization Approach
Supporting Context Resumption in Collaboration
Environments

Liliana Ardissono, Gianni Bosio, and Marino Segnan

Dipartimento di Informatica, Università di Torino, Italy
{liliana,bosio,marino}@di.unito.it,
WWW home page: http://www.di.unito.it/ liliana,
http://www.di.unito.it/ marino

Abstract. In the research on Computer Supported Cooperative Work, activity
awareness is considered a key feature for the coordination of users’ activities.
We propose a model for the visualization of recent activity awareness information
organized at two abstraction levels: (a) the upper level, represented as a tag cloud,
provides a general view of the degree of activity occurred in the user’s collabora-
tions. (b) the lower level is a detailed view of the occurred events, structured on
the basis of the user’s activity contexts. The results of a user study show that the
adoption of the proposed solution is preferable over a standard awareness space
providing a direct access to complete awareness information.

Keywords: activity awareness support, workspace awareness, collaboration environ-
ments, Web 2.0.

1 Introduction
The research on groupware and Computer Supported Cooperative work describes aware-
ness support as a key feature for collaborative environments, in order to enable users
to maintain an up to date view of their collaborations. In particular, [1] introduces the
activity awareness concept to represent “the awareness of project work that supports
group performance in complex tasks”.
Activity awareness support involves notifying the user about many different types of
information, concerning collaborators, artefacts to be manipulated, actions performed
by others, pending tasks, etc. Thus, a major issue to be addressed is that of preventing
the user from being overloaded by an excessive amount of data to be inspected (i) while
(s)he operates in the collaboration environment, and (ii) every time (s)he resumes the
state of an activity context; e.g., after having been out of office for some time.
The risk of overloading users was evident in former collaboration environments;
e.g., see [2]. However, nowadays it is even more problematic, as private and corporate
users are increasingly using online services to carry out their activities by exploiting
the ubiquitous environment offered by the Internet [3, 4]. Therefore, for each user, the
number of private and shared activity contexts to be handled in parallel, and the amount
of awareness information to deal with, are much larger than before.
16 L. Ardissono, G. Bosio, and M. Segnan

In order to support an efficient resumption of the state of the user’s collaborations,
we propose a two-level model for the visualization of recent awareness information
which provides a synthesis of the evolution of the user’s activity contexts, from which
the details of the occurred events can be retrieved on demand. The idea is that of en-
abling the user to quickly understand the degree of activity occurred in her/his collab-
orations in order to decide whether some of them deserves to be inspected in detail.
For this purpose, we have designed the higher visualization level as a tag cloud (the
Awareness Cloud) whose nodes
– represent activity contexts and users, depicting the level of occurred activity in the
selected time interval by means of their relative size in the cloud;
– are direct links to projections of the awareness space handled by the collaboration
environment, focused on specific activity contexts/users. These projections form
the lower visualization level and support a direct access to recent awareness events
from particular perspectives.
These two views complement the thorough information provided by standard awareness
spaces by enabling the user to access information incrementally and in a focused way.
We conducted an experiment with end users to assess how people interacted with
these views. The results revealed that, in terms of improved users’ performance, our
designed Awareness Cloud represents an added value to an awareness space structured
on the basis of the user’s activity contexts because it helps users to quickly access the
information required to answer specific information needs.
In the following, Section 2 presents our visualization model. Section 3 describes the
user study we carried out and discusses its results. Section 4 compares our proposal to
the related work and Section 5 concludes the paper.

2 Presenting Recent Activity Awareness Information
The provision of awareness information is challenging: on the one hand, push tech-
nologies can be employed to notify users about the occurred events, e.g., via Instant
Messages or e-mail. However, they can generate interruptions possibly having a disrup-
tive effect on users’ attention and emotional state [5]. On the other hand, as discussed
in [6], users acknowledge notifications as disruptive, yet opt for them because of their
perceived value in providing awareness. Moreover, as reported in [7], users are ob-
served to frequently switch among different activity contexts, with a consequent effort
in resuming the state of the contexts they enter.
One way to address the trade-off between keeping users up-to-date about the evo-
lution of their collaborations and interrupting them is the provision of an incremental
access to awareness information. In fact, this solution gives a flavor of what has hap-
pened in the users’ activity contexts and supports a quick access to the details they need,
on demand. The visualization model we propose follows this approach and is thus pro-
posed as an awareness layer to be superimposed over a standard awareness space, in
order to provide views on such space, focused on the recent past and on specific infor-
mation needs. As such information is not enough to reconstruct the complete history of
a collaboration, our visualization model assumes the existence of a separate awareness
Activity Awareness Visualization Approach Supporting Context Resumption 17

space presenting the long-term event history. See [8] for a proposal of how such a space
could be organized.

2.1 Context-dependent Management of Awareness Information
For each user of a collaborative environment, the awareness events to be visualized
concern actions performed by her/himself, or by her/his collaborators, while they use
the business services integrated in the environment. In order to support a structured,
context-dependent presentation of information to the user, our visualization model makes
two main assumptions:
– The user’s activity contexts are explicitly modeled, as well as the collaboration
groups associated to such contexts.
– The awareness events generated by the services integrated in the collaboration en-
vironment are classified in their reference activity contexts, so that they can be
managed in a structured awareness space reflecting the user’s collaborations and
private activities.

Fig. 1. Awareness cloud of a user of a collaboration environment (user utntest1@gmail.com).

As described in [9, 10], we represent the user’s private and shared activity contexts
at different granularities, considering the following types of contexts:
– Collaboration sphere: this is a thematic group, similar to a virtual community, used
to keep in touch with each other. For instance, the “family” sphere could be defined
to keep track of the communication concerning the user’s family.
– Activity frame: this is a more or less structured project, which a user can define in
order to collect artefacts of interest around a topic and manage activities aimed at
reaching a goal, possibly in collaboration with other users. For instance, an activity
frame could represent a work project aimed at preparing a conference paper.
– Task: this is used to specify and carry out the execution of an activity, possibly
shared with other users; e.g., writing a section of the above mentioned conference
paper. A task may include artefacts to be manipulated and can have a deadline.
Tasks are created within activity frames and can be related to each other according
to partial order relations, in order to coordinate the execution of complex activities.
18 L. Ardissono, G. Bosio, and M. Segnan

[9] and [10] present a framework for the development of user-centered service clouds
supporting an explicit management of contexts and the consequent classification of
awareness events. The visualization model proposed here builds on that architecture but
could be applied to a different one, as long as it guarantees the association of awareness
events to actors and contexts.

2.2 Two-level Presentation of Awareness Information
We propose to visualize the recent activity awareness information for a user in the
Awareness Cloud. This is a tag cloud which shows the degree of activity occurred in the
user’s private and collaboration contexts during the time interval selected by the user.

Fig. 2. Detailed views on awareness information concerning context “LAVORO A”.

As shown in Figure 1, the Awareness Cloud for a user U is organized as follows:
– The nodes represent four types of entities: user nodes are associated to U ’s col-
laborators; e.g., see node CLAUDIO. Collaboration sphere, activity frame and task
nodes are associated to the user’s collaboration spheres, activity frames and tasks,
respectively. For instance, node LAVORO A represents a collaboration sphere.
– The relative size of each node in the cloud represents the degree of activity in the
selected time window and depends on the number of associated awareness events
that have been collected in the collaboration environment. Specifically, user nodes
visualize the degree of activity of the represented users within U ’s activity contexts,
as the operations performed by users in other contexts cannot and must not be
disclosed to the user. The other types of nodes summarize the degree of activity
occurred in the contexts they refer to.
– The user can specify the starting and end time of the interval for the generation of
the cloud in order to visualize the evolution of her/his activity contexts along time.
Activity Awareness Visualization Approach Supporting Context Resumption 19

Moreover, a “catch up” button enables the user to refresh the cloud by setting the
starting time to the current time. This is useful when the user is not interested in the
recent event history any more.
Thus, for each user, a dynamic awareness cloud is generated, which reflects the activity
contexts (s)he engages in and the selected time interval.

Fig. 3. Architecture of the collaboration environment supporting activity awareness.

Each node of the cloud is linked to a view on the main awareness space of the
collaboration environment which shows the related awareness events in detail. Each
view is indeed a projection of the awareness space, focused on the context represented
by its source node in the Awareness Cloud and on the time interval selected for the
generation of the cloud. Specifically, the page linked to a user node displays the list
of events describing the actions performed by the represented user. Moreover, the page
linked to an activity context node displays the list of events occurred in that context.
Figure 2 shows the event history associated to node LAVORO A of the awareness
cloud: in order to support the navigation of events, these can be sorted by date, ac-
tor, content and task (in fact events are managed as structured objects, with features,
within the collaboration environment. Notice that, if the context associated to a view in-
cludes any nested contexts (in this case, PROGETTO EUROPEO and CONFERENZA
OLTREOCEANO, which are two activity frames defined within the LAVORO A col-
laboration sphere), the page includes links which the user can follow in order to visual-
ize the events concerning such nested contexts. Thus, the projections on the awareness
space are hierarchical.
It should be noticed that the cloud includes a maximum number of 40 elements to
be visualized at each time because, as discussed in [11], a cloud with too many tags
can be puzzling and hard to read. Should more than 40 elements be eligible for visu-
alization, those with least elements would therefore be dropped. The user can however
personalize the cloud by suppressing nodes in order to avoid the visualization of users
20 L. Ardissono, G. Bosio, and M. Segnan

and/or activity contexts (s)he is not interested in. Moreover, we are extending the cloud
generation model in order to allow the user to specify high-priority nodes, associated
to users and/or contexts which the user wants to monitor with particular attention. Such
nodes will not be dropped from the cloud and will be depicted in a different color for
easy identification.
We integrated our visualization model in a collaboration environment developed by
exploiting the Personal Cloud Platform (PCP) [9], which supports the development of
customized collaboration environments by integrating heterogeneous software compo-
nents in order to answer specific functional needs. Figure 3 shows the overall system
architecture and highlights the generation of the Web pages according to our proposed
model (integrated in the Awareness Support Tool of the environment). The PCP enables
the user to specify her/his collaboration spheres and to synchronize heterogeneous busi-
ness services accordingly. Moreover, it offers the Collaborative Task Manager (CTM,
[10]) for the management of activity frames and tasks and for the classification of the
awareness events generated by the user’s actions in the related activity contexts. The
CTM offers a User Interface which enables the user to interact with business services
(e.g., to create or manipulate artefacts) within a specific activity context and to classify
awareness events in the appropriate activity contexts.

3 Tests
3.1 Description
We conducted an experiment to evaluate the impact of the introduction of the Aware-
ness Cloud on users’ experience. We wanted to test a hypothesized causal relationship
between the introduction of the Awareness Cloud on top of an awareness space struc-
tured on the basis of the user’s activity contexts (henceforth, context-aware awareness
space) and people’s performance during a task.
Our research question was “Does a context-dependent tag cloud modify the level of
performance of the users with an activity awareness space?”. If the answer is positive,
which case can give best results?
Hypothesis (Ha): The introduction of a custom tag cloud to enhance a context-dependent
activity awareness space (i.e., an awareness space structured on the basis of the user’s
activity contexts) will improve users’ performance on an awareness information seeking
task, in terms of execution times and number of errors.
Sixteen volunteers participated as participants in this experiment (10 men and 6
women). All participants were students or staff within the University of Torino and
performed the test for free, without any reward.
The experiment had a single-factor, between-subjects design. Two treatments were
applied - one experimental treatment and one base-case control treatment. The exper-
imental treatment consisted in a context-dependent activity awareness space enhanced
with an Awareness Cloud, while the context-dependent awareness space alone was re-
puted as the base-case.
Each treatment condition was considered as an independent variable. Participants’
performance was considered as a dependent variable and was calculated considering
Activity Awareness Visualization Approach Supporting Context Resumption 21

two objective measures: number of committed errors and time needed to complete the
task. Participants were divided into two groups of eight people, and each group received
one single treatment. Such design was aimed at preventing side effects such as practice
and fatigue. Users were also given two questionnaires: one before the task, the other
after task completion. The first questionnaire was meant to evaluate users’ background
about collaborative applications. The second questionnaire was meant to evaluate users’
opinion on the adopted User Interface solution.
The experimental task was designed as an information recovering and comprehen-
sion one, simulating a typical, asynchronous reception of awareness information in a
collaboration environment. Users in both groups were briefed about their scenario be-
fore the beginning of the task: as participants of three different collaboration groups,
they had received awareness information regarding other users’ activities, that was still
to be read. Such information was collected in a structured list (the activity awareness
space), where each event-related element was organized on the basis of its originating
activity context; the recent activity awareness consisted of 13 events. Users were then
asked to answer six questions, whose answers could be found by navigating the events.
Questions 1, 2 and 4 were general, quantitative oriented ones, such as “who is the most
active user in a certain task”; questions 3, 5 and 6 were more specific, as for example
“list every task and collaboration sphere a certain user is involved in”.
All participants used the same activity awareness space for this purpose. The only
difference between the two treatments was the presence (or the absence) of the Aware-
ness Cloud, combined with the visualization of recent events. Users within the exper-
imental treatment group could therefore access particular “projections” of the aware-
ness space by clicking on the corresponding node of the Awareness Cloud. Each user
was also given information about the nature of their (simulated) collaborations, such
as names of collaboration groups, projects, tasks and involved users. Such instructions
were available to participants as a reference throughout the whole experiment. Each
participant was engaged in testing activity for a period of about 15 minutes.

3.2 Results and Discussion

Fig. 4. Test results.
22 L. Ardissono, G. Bosio, and M. Segnan

The first questionnaire was aimed at measuring the level of practice each user had
with collaborative applications, within two different environments: workplace and pri-
vate life. Each user could choose between four non-decreasing levels of practice, rang-
ing from 0 to 3. Results showed no significant difference in the level of practice each
user had with collaborative applications in both environments.
Figure 4 shows the results of the user tests. The figure is split in two parts, and both
parts show mean values for the two treatments: the values are referred to the number of
errors and execution times (in seconds) respectively, and are defined on the y-axes; the
x-axes identify the treatment group.
We used an unpaired Welch’s t-test (which does not assume equal variances) to an-
alyze collected data. An alpha level of 0.05 was used to make decisions of significance.
We found a significant effect either for number of errors (t = -2.38, p = 0.049 < 0.05)
and for execution times (t=-3.15, p = 0.011), that lead us to reject null hypothesis of no
difference between the treatments, and to accept our hypothesis: the introduction of the
Awareness Cloud lead to an improvement of users’ performances.
In the second questionnaire we asked users to evaluate their own experience with
the User Interface they operated with: awareness space plus Awareness Cloud for ex-
perimental group, awareness space alone for control group. Each user could choose
between seven non-decreasing levels of satisfaction, ranging from 0 to 6. The exper-
imental group expressed a mean value of 5,81 for their UI (st.dev = 0,55), while the
control group expressed a mean value of 5,31 (st.dev = 1,02).
The results of this experiment revealed that the introduction of the Awareness Cloud
significantly improved users’ performance, in terms of times of execution and number
of errors. First-hand observations of participants behavior in this experiment lead us to
grasp two aspects that may explain these results:

– The Awareness Cloud proved itself as very easy to understand and to use, and
showed a good level of integration with the awareness space. Indeed, the users of
the experimental group were left free to choose arbitrarily whether to adopt it or
not, but every one of them (even those who did not know what a tag cloud was)
opted for its use since the first question.
– The Awareness Cloud allowed users to express fast and precise queries by clicking
on the desired nodes, with a User Interface that was valued as “practical, good and
interesting”. Navigating into the awareness space in isolation did not prove itself
as immediate and error-proof as the Cloud: users of the control group who did not
commit errors tended to spend more time doing their tasks, probably due to the
need of verifying their choices with more accuracy.

Users indicated as a major drawback of the Awareness Cloud the fact that it made
hard to spot nodes with a very low density of events: while it could be much faster
to identify high density elements (specifically, groups of users and highly active tasks),
those written with the smallest font (such as low activity tasks) might get lost among the
crowd. This aspect is typical of a tag cloud [11] but could be addressed by supporting
a personalized configuration of the cloud, based on the user’s interests. Specifically, we
plan to enable the user to configure the Awareness Cloud by specifying which elements
(s)he wants to monitor with most attention. When the cloud is generated, such elements
Activity Awareness Visualization Approach Supporting Context Resumption 23

will then be displayed with a different color (e.g., red instead of traditional light blue)
and would never be omitted when the cloud is too large.

4 Related Work

Most groupware and project management tools only offer standard awareness spaces
which show the list of occurred events organized by collaboration group; e.g., BSCW
[12]. Other systems, such as CANS [13], support the presentation of awareness events
in different formats (such as lists and tables), but events are classified by group/shared
directory. Furthermore, [14] proposes a radar view of awareness events, which are only
classified by source application.
In [2], AwarenessMaps are proposed to provide the members of shared workspaces
with an overview of users and documents: “the PeopleMap shows an array of pictures of
active users fading out over time; and the DocumentMap provides a schematic overview
of the structure of a shared workspace and indicates recent changes.” Moreover, [15] in-
troduces a pictorial representation of incoming e-mails (Info-Lotus), divided in groups
and sub-groups in order to represent conversation threads. Our proposal makes a step
forward in this direction by visualizing the recent awareness information at different
granularity and abstraction levels. The granularity aspect concerns the generality of
the activity context to be considered and is motivated by the fact that users engage in
different types of collaborations, such as thematic groups (e.g., small or large virtual
communities), more or less structured projects, and specific tasks. The abstraction as-
pect enables the user to receive a synthesis of the evolution of her/his activity contexts
and to select the contexts to be inspected in detail.
Recently, the research about collaboration in online communities has focused on
activity awareness in order to inform users about who is active in the topics of interest of
the community, which kind of contribution has been provided, and similar. For instance,
[16] proposes a “star” view of users, aimed at showing their degree of activity in the
community. Moreover, [17] proposes a visualization of activity awareness in CiteULike,
which exploits radial time bands to show the time period during which the user/group
activity (or the activity on a topic) has occurred. Our proposal differs from those works
because, besides modeling individual users and groups, we model the user’s activity
contexts. Specifically, the visualization we propose enables the user to assess the state
of her/his collaborations or to focus on aspects, such as a particular task.

5 Conclusions

This paper has described a visualization model supporting the incremental access to
activity awareness information in a collaboration environment. Our model presents the
awareness information at different levels of detail in order to provide the user with
a general view on what has recently happened in her/his collaborations, and enable
her/him to retrieve detailed information on specific activity contexts. A user study
showed that the adoption of the proposed solution is preferable over a standard aware-
ness space providing a direct access to complete awareness information.
24 L. Ardissono, G. Bosio, and M. Segnan

Before concluding, it is worth mentioning that the model presented in this paper is
the first step towards the development of an adaptive awareness support service enabling
users to receive a personalized view of the information they need, depending on their
interests and activities. In fact, the current model for the generation of the Awareness
Cloud is only based on the user’s activity contexts and on the selected time interval for
the visualization of information. Personalized clouds could be generated by enabling
the user to explicitly select “high-priority” contexts (as proposed in Section 2), but also
by tracking the user’s interests across activity spaces along time, and by dynamically
configuring the Cloud in order to focus it on the most relevant ones; e.g., see [18] for
a similar approach applied to notification management. In our future work, we plan to
extend our awareness model towards the provision of adaptive workspaces which tailor
both the presentation of information (e.g., awareness information) and their services to
the dynamics of the collaboration activities carried out by users; e.g., see [19].

References

1. Carroll, J., Neale, D., Isenhour, P., Rosson, M., Scott McCrickard, D.: Notification and
awareness: synchronizing task-oriented collaborative activity. International Journal of
Human-Computer Studies 58(5) (2003) 605–632
2. Gross, T., Wirsam, W., Gräther, W.: AwarenessMaps: Visualising awareness in shared
workspaces. In: Extended Abstracts of the Conference on Human Factors in Computing
Systems - CHI 2003. ACM, New York (2003) 784–785
3. V.S. Pendyala, S.S.Y. Shim: The Web as the ubiquitous computer. Computing now (Septem-
ber) (2009) 90–92
4. Prinz, W., Löh, H., Pallot, M., Schaffers, H., Skarmeta, A., Decker, S.: ECOSPACE - towards
an integrated collaboration space for eProfessionals. In: Proc. of 2nd Int. Conf on Collabora-
tive Computing: networking, applications and worksharing, Atlanta, Georgia (2006) 39–45
5. Bailey, B., Konstan, J., Carlis, J.: The effects of interruptions on task performance, annoy-
ance, and anxiety in the user interface. In: Proc. INTERACT’01, Tokyo (2001) 593–601
6. Iqbal, S., Horvitz, E.: Notifications and awareness: a field study of alert usage and pref-
erences. In: Proc. of the 2010 ACM conference on computer supported cooperative work
(CSCW 2010), Savannah. Georgia (2010) 27–30
7. Mark, G., Su, N.: Considering Web 2.0 technologies within an ecology of collaborations.
In: Proc. of SociUM: Adaptation and Personalisation in Social Systems: Groups, Teams,
Communities, Corfu, Greece (2007) 50–59
8. Ardissono, L., Bosio, G., Goy, A., Petrone, G., Segnan, M.: Managing context-dependent
workspace awareness in an e-collaboration environment. In: Proc. of WI/IAT09 workshop
”Intelligent Web Interaction” (IWI 2009), Milano, Italy, IEEE (2009) 42–45
9. Ardissono, L., Goy, A., Petrone, G., Segnan, M.: From service clouds to user-centric personal
clouds. In: Proc. of IEEE 2009 International Conference on Cloud Computing (CLOUD-II
2009), Bangalore, India, IEEE (2009) 1–8
10. Ardissono, L., Bosio, G., Goy, A., Petrone, G., Segnan, M.: Open, collaborative task man-
agement in web 2.0. In: Proc. of MCCIS 2010 IADIS multiconference on computer science
and information systems, Freibug, Germany, IADIS Press (2010) 20–27
11. Bateman, S., Gutwin, C., Nacenta, M.: Seeing things in the clouds: the effect of visual
features on tag cloud selections. In: Proc. of 19th ACM Conference on Hypertext and Hy-
permedia (HT’08), Pittsburgh, PA, USA (2008) 193–202
Activity Awareness Visualization Approach Supporting Context Resumption 25

12. Horstmann, T., Bentley, R.: Distributed authoring on the Web with the BSCW shared
workspace system. StandardView 5(1) (1997) 9–16
13. Amelung, C., Laffey, J., Turner, P.: Supporting collaborative learning in online higher edu-
cation through activity awareness. In: Proc. 8th Int. Conf. on Computer Supported Collabo-
rative Learning (CSCL ’07), New Brunswick, New Jersey (2007) 75–77
14. van Dantzich, M., Robbins, D., Horvitz, E., Czerwinski, M.: Scope: providing awareness of
multiple notifications at a glance. In: Proceedings of the Working Conference on Advanced
Visual Interfaces, New York (2002)
15. Zhang, L., PhD, N., Vronay, D.: Info-Lotus: a peripheral visualization for email notification.
In: Proc. of CHI 2005, Portland, Oregon (2005) 1901–1904
16. Vassileva, J., Sun, L.: Using community visualization to stimulate participation in online
communities. e-Service Journal 6(1) (2007) 3–39
17. Baishya, D., Brusilovksy, P.: CiteAware: visual group awareness for a reference sharing sys-
tem. In: Proc. of workshop on visual interfaces to the social and semantic web (VISSW2009),
Sanibel Island, Florida (2009) 153–156
18. Wang, Y., Gräther, W., Prinz, W.: Suitable notification intensity: the dynamic awareness
system. In: GROUP’07: Proc. of the 2007 Int. ACM Conference on supporting group work,
New York, NY (2007) 99–106
19. Veiel, D., Haake, J., Lukosch, S.: Facilitating team-based adaptation of shared workspaces.
In: Proc. of 2010 Int. Symposium on Collaborative Technologies and Systems (CTS),
Chicago, IL (2010) 275–284