=Paper=
{{Paper
|id=Vol-1776/paper12
|storemode=property
|title=SmartFit: Lifelogging for Teams of non-Professional Athletes
|pdfUrl=https://ceur-ws.org/Vol-1776/paper12.pdf
|volume=Vol-1776
|authors=Stefano Valtolina,Barbara Rita Barricelli
|dblpUrl=https://dblp.org/rec/conf/nordichi/ValtolinaB16
}}
==SmartFit: Lifelogging for Teams of non-Professional Athletes==
https://ceur-ws.org/Vol-1776/paper12.pdf
SmartFit: Lifelogging for Teams of non-Professional
Athletes
Stefano Valtolina, Barbara Rita Barricelli
Department of Computer Science, Università degli Studi di Milano
Via Comelico 39/41, 20135 Milano, Italy
{valtolina, barricelli}@di.unimi.it
Abstract. In IoT domain, communities of domain experts, having different skills
in specific areas of endeavor, need an effective and easy-to-use strategy for man-
aging physical devices and their data streams. Specifically, in this paper we are
interested in discussing a sociotechnical study aimed at designing an IoT ecosys-
tem to be used by non-professional sport teams. Nowadays, coaches and trainers
want to take advantage from the plethora of sensors and applications that can be
used for monitoring the quality of their athletes’ activities and behaviors, but they
need an easy to use environment to perform proper policies, and the visualization
and analysis of relevant events.
Starting from a definition of End-User Development (EUD) designed around the
pervasive requirements of IoT applications, we describe a system able to provide
coaches and trainers with a tool for monitoring, studying the flow of events to
detect significant situations, and triggering proper actions through the use of fil-
tering and temporal operators.
Keywords: Internet of Things, EUD, Rule Editor.
1 Introduction
The Internet of Things (IoT) concept is spreading thanks to the evolution of sensor
technology and its use that is becoming more and more mobile and pervasive [1].
Tech-nology is part of everyday life, and some of this technology plays a role in
sports prac-tice. New wearable trackers to place on wrists, ankles, or waistbands
are driving the sport practice towards revolutionary changes together with other
accessories or mobile applications that can be used for monitoring physical
activities, quality of sleep, and diet. This type of integration is what characterizes
the so-called lifelogging: keeping track of the collected data through all the
everyday or occasional activities that may influence people’s quality of life. Sports
technology can exploit this Quantified-Self trend for helping people to adopt a
healthy lifestyle or for supporting coachers to inves-tigate innovative strategies of
training aimed at monitoring their athletes' physical ac-tivities and performances.
The information this sport equipment can record — from steps taken to sleep hours
— can be gathered and shared for tracking the quality of the sport activities or the
quality of life itself of the athletes.
76
Proc. of Fourth International Workshop on Cultures of Participation in the Digital Age - CoPDA 2016
Gothenburg (Sweden), October 23, 2016 (published at http://ceur-ws.org).
Copyright © 2016 for the individual papers by the papers' authors. Copying permitted for private and academic purposes.
This volume is published and copyrighted by its editors.
One of the most exciting trends in the fitness industry is the emergence of the online
fitness coach, web-based personal training services designed to give you the guidance
you need to reach your goals; nevertheless, what we want to investigate in this paper is
the problem seen from the opposite side. How coaches and trainers can use the data
coming from the sport trackers or other tracking applications for monitoring their ath-
letes. To connect in a network everyday devices and applications allows coaches and
trainers to receive data, and at the same time to perform comparisons among athletes,
but also understand when critical or interesting situation happen for each individual.
Programming the behaviors of devices and applications is becoming suitable for eve-
ryone thanks to the design of interfaces that support customization, personalization, and
to some extent also End-User Development (EUD) [2, 3]. To design and create new
applications, it is less and less necessary to learn actual programming languages or to
have previous software coding experiences. The peculiar structure of non-professional
sport organizations is characterized by the existence of small teams with athletes who
live different kind of lives, being professionals in different domains, and meeting only
for some hours a week. Keeping track of their habits, in terms of physical activity,
nutrition, sleep and so on, would help the coaches in understanding the variety of the
team members and finding successful schemes of training. For managing such applica-
tion domain, we need to define a EUD strategy for supporting domain experts in defin-
ing business polices and rules for detecting relevant and critical events.
2 Design Model
The knowledge associated with the design of the highly dynamic data processing that
characterizes an IoT system is tacitly distributed among the various design communi-
ties. Specifically, in lifelogging and quantified-self applied to the management of non-
professional sport teams, the communities are: IoT engineers, Coaches and Trainers,
and Athletes. To perform their activities, coaches and trainers does not want to be bored
by dealing with technical aspects that concern how to connect, maintain, and set up the
devices and sensors to be used by the IoT ecosystem. These activities need to be carried
out by IoT sensors/devices engineers whereas the role of coaches and trainers is to col-
laborate in guiding, instructing, and training the members of a sport team. To exploit at
best, the potentials of IoT in their practice requires a specialized interactive system for
designing rules for defining what actions have to be performed in response to specific
events. They act as End-User Developers by designing the rules to be used to supervise
athletes’ performances and lifestyle and they also analyze the gathered data in their
interactive system. Several solutions have been proposed to bridge the communication
gap and to design usable interactive systems [4], [5]. Our design strategy stems from
these works and aims at supporting multidisciplinary design teams’ collaboration and
to foster their situated innovation by means of several EUD methods. Our model fol-
lows a bottom-up approach that breaks down static social structures to support richer
ecologies of participation. It offers three different levels of participation and design
activities: i) Dataflow design level; ii) Rules design level; iii) Rules deployment, where
the rules are deployed and end users use the environments and tools. At dataflow design
77
Proc. of Fourth International Workshop on Cultures of Participation in the Digital Age - CoPDA 2016
Gothenburg (Sweden), October 23, 2016 (published at http://ceur-ws.org).
Copyright © 2016 for the individual papers by the papers' authors. Copying permitted for private and academic purposes.
This volume is published and copyrighted by its editors.
level, IoT Engineers need to configure the network of sensors and services for manag-
ing the data-flow to be served at the rules design level. The outcome of this environment
is the detection of a set of relevant events that coaches and trainers need to manipulate
for monitoring the physical activities or daily behavior of their athletes. We identify an
event with its temporal, spatial, and thematic dimensions that can be exploited both for
the identification of the useful information needed to face a given event and for the
analysis and forecast of useful activities to be notified to the users. In our architecture,
the dataflow design environment is a Web environment where IoT engineers can drag
and drop different sensor data sources and visually apply on them a set of operations.
This application offers an engine and graphical environment for data transformation
and mashup. Further details are reported in [6]. In what follows, we focus on the second
environment, the Rules editor, which aim is to offer a graphical visual environment for
supporting coaches and trainers to monitor a flow of events detected at the first level of
our architecture, for analyzing actions and behaviors of their teams of non-professional
athletes.
3 Rule Editor Interface
Visual strategies typically used in IoT field for modelling Event-Condition-Action rules
can be described through the most famous systems that apply them: IFTTT 1 and
Atooma 2. These applications allow users to define sets of desired behaviors in response
to specific events. This is made mainly through rules definition-wizards that rely on the
sensors/devices states. The visual strategy aims at creating automated rules by using
graphical notation for programming statements such as: “IF this DO That” or “WHEN
trigger THEN action”. However, the language is not expressive enough for the specifi-
cation of more sophisticated rules based on time and space conditions. For example, in
a scenario where we want to monitor data related to health conditions and behavior of
a group of athletes, apart from supporting logical combinations of conditions, we need
to specify timing comparison between different events. For example that, the athlete
has taken less calories of those burned in a physical activity that has been taken place
after lunch. A second type of applications stems from another outstanding work done
with Yahoo's Pipes 3. Such applications offer solutions based on graphical environ-
ments for data transformation and mashup. The idea is to provide a visual pipeline gen-
erator for supporting end users in creating aggregation, filtering, and porting of data
originated by sources. The visual strategies adopted by such Yahoos Pipes-compliant
solutions are promising techniques but, in our opinion, they present some lacks. Some
studies [7] also found that, although these strategies tried to simplify mashup develop-
ment, they are still difficult to use by non-technical users, who encounter difficulties
with the adopted composition languages [8].
Our Rule Editor user interface leverages the issues required for expressing complex
conditions leading to a system that can be easily used by non-expert users. As depicted
1 https://ifttt.com
2 http://www.atooma.com
3 https://pipes.yahoo.com/pipes
78
Proc. of Fourth International Workshop on Cultures of Participation in the Digital Age - CoPDA 2016
Gothenburg (Sweden), October 23, 2016 (published at http://ceur-ws.org).
Copyright © 2016 for the individual papers by the papers' authors. Copying permitted for private and academic purposes.
This volume is published and copyrighted by its editors.
in Figures 2 the interface is based on select widgets that are populated by using the
attributes that characterize the JSON of the flow of events produced by the IoT Engi-
neers. Through a visual notation, domain experts can specify conditions and temporal
operations for implementing the business rules that characterize their activities.
An example of Composite Rule creation is given in Figure 2 where the user has
defined four rules on the base of these conditions: (i) hours of sleep less than 7 (ii)
calories intake at dinner greater than 1,500 (iii) number of steps less than 8,000 (iv)
activity duration less than 45 (minutes). According to these conditions a set of four
events are defined that can be built in this way: E1 AND E2 AND (E3 OR E4). The
meaning of the Composite Rule created in this example is: “if the hours of sleep of the
day before are less than 7 AND if the calories intake at dinner (before the sleep) is
greater than 1,500 AND (if the number of steps is less than 8,000 at day OR the duration
of physical activity is not less of 45 minutes at day THEN send the athlete a message
that warns about the behavior and performances.
The Rule Editor aims at allowing non-technical people to specify rules by using sim-
ple drop-down menus. The conditions can be composed by combining groups of state-
ments connected by using the AND/OR operators. The order of the conditions can be
changed by the user just by dragging and dropping the statements into the right position.
Domain experts can filter data on a certain period of time set by using the “validity
interval” parameter (see Figure 2). Temporal conditions are defined using the automat-
ically assigned names of the Rules as elements to be composed (e.g., E1, E2, E3). An
example of complex temporal condition can be: E4 (activity duration less than 45
minutes) starts from 5 to 10 minutes before of the E2 (Calories intake at dinner greater
than 1,500) and ends from 3 to 7 minutes after E2 ends. In other words, the trainer
wants to check how much her/his athletes eat, if they eat seated at a table or when are
on the move and in how much time. Once a rule is created, it is stored in a repository
for further re-use or for sharing it among members of a community of trainers and
coaches.
4 Ongoing Research
After a first evaluation performed involving HCI expersts, we are now setting up a user
test evaluation for our eWellness system thanks to collaboration with the Centro Spor-
tivo Italiano (Italian Sport Centre). A group of sport teams, all domain experts, will be
involved in order to perform a set of activities concerning the design of data-flows and
related rules from monitoring the athletes’ activities. What we want to study is how far
our approach is able to offer new possibilities both at the design and use time and to
understand how the idea to combine the design and end users’ environments appears to
be successful and effective solution both for domain and technical experts.
79
Proc. of Fourth International Workshop on Cultures of Participation in the Digital Age - CoPDA 2016
Gothenburg (Sweden), October 23, 2016 (published at http://ceur-ws.org).
Copyright © 2016 for the individual papers by the papers' authors. Copying permitted for private and academic purposes.
This volume is published and copyrighted by its editors.
Fig. 1. Example of the composition of a set of rules.
5 References
1. Ashton, K.: That ‘Internet of Things’ Thing. RFID Journal, June (2009). Available online:
http://www.rfidjournal.com/articles/view?4986 (accessed on January 19th, 2015).
2. Lieberman, H., Paternò, F., Klann, M., Wulf, V.: End-User Development: An Emerging
Paradigm. In: H. Lieberman, F. Paternò, & V. Wulf (Eds.) End-User Development, pp. 1--
8). Springer (2006).
3. Barricelli, B.R., Valtolina, S.: Designing for End-User Development in the Internet of
Things. End-User Development. Lecture Notes in Computer Science, 9083, pp. 9-24 (2015).
4. Valtolina, S., Barricelli, B. R., & Dittrich, Y. (2012). Participatory knowledge-management
design: A semiotic approach. Journal of Visual Languages & Computing, 23(2), 103-115.
5. Costabile, M.F., Fogli, D., Mussio, P., Piccinno, A.: End-user development: the software
shaping workshop approach. pp. 183–205, Springer (2006).
6. Mesiti, M., Ferrari, L., Valtolina, S., Licari, G., Galliani, G. L., Dao, M., Zettsu, K. Stream-
loader: An event-driven ETL system for the on-line processing of heterogeneous sensor
data., in: EDBT 2016, Bordeaux, France, March 15-16, 2016., 628-631.
7. Namoun, A., Nestler, T., De Angeli, A. Conceptual and usability issues in the composable
web of software services, in: International Conference on Web Engineering, Springer, 2010,
396-407.
8. Casati, F. How end-user development will save composition technologies from their contin-
uing failures, in: International Symposium on End User Development, Springer, 2011, pp.
4-6.
80
Proc. of Fourth International Workshop on Cultures of Participation in the Digital Age - CoPDA 2016
Gothenburg (Sweden), October 23, 2016 (published at http://ceur-ws.org).
Copyright © 2016 for the individual papers by the papers' authors. Copying permitted for private and academic purposes.
This volume is published and copyrighted by its editors.