978-1-7281-3637-0/19/$31.00 ©2019 IEEE
Dashboard for the VISIR remote lab
Javier García-Zubía
Facultad de Ingeniería
University of Deusto
Bilbao, Spain
zubia@deusto.es
Unai Hernández-Jayo
Facultad de Ingeniería
University of Deusto
Bilbao, Spain
unai.hernandez@deusto.es
Pablo Orduña
LabsLand S.L.
Bilbao, Spain
pablo@labsland.es
Jordi Cuadros
IQS
University Ramón Llull
Barcelona, Spain
jordi.cuadros@iqs.url.edu
Ignacio Angulo-Martínez
Facultad de Ingeniería
University of Deusto
Bilbao, Spain
ignacio.angulo@deusto.es
Gustavo Alves
Inst. Sup. Engenehria de Porto
Inst. Politecnico de Porto
Porto, Portugal
gca@isep.ipp.pt
Vanessa Serrano
IQS
University Ramón Llull
Barcelona, Spain
vanessa.serrano@iqs.url.edu
Aitor Villar
Facultad de Ingeniería
University of Deusto
Bilbao, Spain
aitor.v@deusto.es
Abstract— The VISIR dashboard (VISIR-DB) is a learning
analytics tool connected with the VISIR remote lab. In VISIR,
every action performed by a student from the interface over
the remote laboratory and back is logged and recorded.
VISIR-DB helps visualizing, in a fast and deep way, the
recorded logs from this communication. Using this tool, a
teacher can analyze and understand better how the students
are using the remote lab during their learning process on
analog electronics. With this information, the VISIR platform
can be improved and the use of remote labs can be better
understood.
Keywords— remote laboratory, learning analytics, dashboard
I. INTRODUCTION
VISIR [1]is the most popular and well-known remote lab
(Fig. 1). It is being used by more than fifteen universities and
schools in the world and it is oriented to teach analog
electronics.
Previous work [2-3] reported that the use of VISIR with
DC circuits has a positive learning effect on the students,
according to survey-based studies.
Fig. 1. VISIR remote lab: (left) interface, (right) hardware. Arrows
represent information going back and forth.
Noticeably, when the students are using the VISIR
remote lab under the WebLab-Deusto platform, every action
performed by them is recorded, i.e. what circuit was created
and what measurements were requested (if any). This is a
huge amount of information that can give us a deep
knowledge about how the students are using the VISIR
remote lab.
Currently we work on using this information to get
deeper insights of what happens on an academic task
performed with the VISIR remote lab. This is the goal
behind the VISIR-DB, a learning analytics tool currently in
development [4].
Now we present, using this approach, some details on the
students' use of VISIR in a DC circuits experimental activity.
The objectives of this contribution are then two-fold: (1) to
The authors acknowledge the support provided by the European
Project PILAR. Platform Integration of Laboratories based on the
Architecture of visiR - Erasmus+ Strategic Partnership nº 2016-1-ES01-
KA203-025327.
2019 5th Experiment@ International Conference (exp.at’19)
June 12th – 14th, 2019, University of Madeira, Funchal, Madeira, Portugal
42
Authorized licensed use limited to: b-on: Instituto Politecnico do Porto. Downloaded on February 08,2021 at 16:41:59 UTC from IEEE Xplore. Restrictions apply.
present the use of the VISIR-DB tool in real academic task,
and (2) to discuss what insights are gained from its use.
The paper is focused on the use of the VISIR remote lab
in a Physics course during the second semester of the first
academic year of the double degree in Informatics +
Business Administration at the campus of San Sebastián of
the University of Deusto (Spain). In the 2017-2018 academic
year there were 19 students enrolled.
Physics course is divided in three parts: fundamentals of
electromagnetism, AC and DC circuits and CMOS digital
circuits. VISIR is used when teaching AC and DC circuits. In
DC circuits, the students learn what a simple DC circuit (a
combination of resistors plus one DC power source) is, how
to solve it analytically and how to build and perform
measurements on it.
In a common academic assigment, after solving a given
DC circuit using Ohm’s and Kirchhoff’s Laws, each student
must implement and measure the given circuit to compare
whether the obtained measurements match the analytical
results. Instead of using a hands-on lab, we use the VISIR
remote lab. So, the students build and measure different
circuits using only the VISIR remote lab. The DC circuits
had a maximum of four resistors combined in parallel and/or
series. This topic was covered during 2 weeks at the
beginning of the semester, in February 2018.
In the paper, first the VISIR-DB tool is introduced
(Section II), and then, the results obtained for the single
presented activity are shown and discussed (Section III). At
last, conclusions and future work are stated in Section IV.
II. VISIR-DB DESCRIPTION
VISIR-DB has been designed and developed as a Shiny
application [5]. The Shiny app we are presenting is arranged
in four sections: Data Input, Global Results, Circuit-based
Analysis, and User-specific Results. Each section is
organized into several subsections and panels, in which
meaningful visualizations are displayed.
The Data Input section allows selecting the file to be
analyzed. A CSV corresponding to the recording of the
interactions in the WebLab-Deusto platform is expected.
Summary data as number of users or date range are provided
for double-checking purposes.
The Global Results section provides different summaries
and visualization intended to provide a summarized view of
the group of students, both in terms of time and amount of
work (number of actions and experiments performed).
Attending to the circuits, and not to the students, the
Circuits-based Analysis section offers the list of circuits
implemented by the users and shows when and who built
each of the circuits. The User-specific Results Section
facilitates information for a specific user and for the circuits
created by him.
Next, we will show and analyze the results given by
VISIR-DB for the 19 students described above.
III. ANALYSIS OF A TASK WITH THE VISIR-DB
The VISIR activity took place in two classroom sessions
of two hours, so four hours in total in the classroom were
with VISIR. Two other sessions of two hours were devoted
to the analytical solving of the electronics circuits using
Ohm’s and Kirchhoff’s Laws.
During the first session the students had the first contact
with VISIR and after some examples presented by the
teacher to the students they learnt how to create an
electronic circuit with only resistors, without power source.
Then the students learnt how to measure the total resistance
of a circuit. After some examples made by the teacher with
the students, they had to create and measure 20 new
electronic circuits with two resistors of 1 k and two
resistors of 10 k. A google excel was shared by the
teacher, and the students had to write there the obtained
measurements. Using the google excel the concept of
measurement error was introduced. After the session, the
students were asked to measure only the total resistance of
all the given circuits, as homework.
During the second session the students started to power
the circuits and to measure the voltage and current. In a first
step the teacher showed the students how to measure the
voltage in a powered circuit. Then, the current measurement
was introduced to the students. In this case more time was
devoted to this because current measuring is harder than
voltage measuring. After some examples made all together,
students and teacher, the students started to work alone
measuring voltages and currents of a set of given circuits.
After the session, the students were asked to measure all the
given circuits as a home work.
During the two sessions the teacher showed the students
the most common errors made when creating, powering and
measuring a circuit. Some errors can be defined as
instrumental or manual errors, but other errors must be
considered conceptual errors.
All the actions made by the students when they were
mounting and measuring circuits either at home or in the
classroom were recorded by WebLab-Deusto in a trace file.
To analyze the trace file, such as pulled from the system
database, the first step, see Fig. 2, consists on loading the
data to VISIR-DB. Once the file is loaded, the number of
users, 19 in this case, and the date range, from February, 7
th
to February, 24
th
, are shown.
We move then to analyzing the global results for the
class. Looking at Fig. 3 we can see that the average use time
on task of VISIR was 4.25 hours per student, with a clearly
right-tailed distribution. The minimum was less than two
hours, the maximum was 9.64 hours and the total was 80.81
hours.
Fig. 2. Data loading from Weblab-Deusto RLMS
2019 5th Experiment@ International Conference (exp.at’19)
June 12th – 14th, 2019, University of Madeira, Funchal, Madeira, Portugal
43
Authorized licensed use limited to: b-on: Instituto Politecnico do Porto. Downloaded on February 08,2021 at 16:41:59 UTC from IEEE Xplore. Restrictions apply.
Fig. 3. Histogram of time on task, using VISIR, per student
Fig. 4. Temporary distribution of the students work; rows are dates and
columns are users, the color scale indicates time on task.
Fig. 4 shows hours of work of each student day by day.
Yellow indicates that the activity was low, and blue remarks
that the activity was high. Attending to this graphic the
teachers can see the activity of each student. The instructor
can also get insights on the time spend using the remote lab
off-class by each user. In this case, students 3, 5 and 13 are
the ones that seem to have put more effort into this task.
Looking at Fig. 5 we can analyze the VISIR platform
attending to the number of performed circuits. 5077 circuits
were created by the students, the mean was 267 circuits per
student, the minimum number of circuits created by one
student was 70, and the maximum was 653. We can see that
there is a huge difference among the students. Like the time
distribution, the number of circuits is a right-tailed
distribution, with some users performing a quite large
number of experiments.
Fig. 5. Histogram of the number of circuits per user
Fig. 6. Two equivalent DC circuits
It is important to remark that in this first analysis in the
VISIR-DB the two circuits of Fig. 6 are taken as different,
when they are equivalent. To account for this, the VISIR-DB
translates any circuit into an equivalent, normalized, one. If
two circuits have the same components and connections
(including the multimeter), then their normalized circuits are
equal. The normalized circuits represent then different
measurement made in a circuit.
Attending to the previous statement, VISIR-DB refines
the initial analysis with the Fig. 7. Now we can see that the
number of different circuits is 1004, the mean of circuits is
95 per student, one student created only 47 different circuits,
and another created 181, almost four times more. We can
also read than 3 students implemented a larger number of
different experiments than the rest of the class.
Fig. 7. Histogram of the number of circuits per user
2019 5th Experiment@ International Conference (exp.at’19)
June 12th – 14th, 2019, University of Madeira, Funchal, Madeira, Portugal
44
Authorized licensed use limited to: b-on: Instituto Politecnico do Porto. Downloaded on February 08,2021 at 16:41:59 UTC from IEEE Xplore. Restrictions apply.
Fig. 8. Students activity; number of normalized circuits in front of time on task.
VISIR-DB also allows us to analyze the activity of the
students. Fig. 8 shows students work (number of normalized
circuits) in front of time on task (time spent using VISIR).
This chart makes easy to spot students that may have
difficulties with task or the class. Why are some students
spending so little time to the task and building such a small
number of different circuits? Are they getting confused? In
our case study, three hard-working students are easily
observed and a handful of students in the bottom part of the
graph that may benefit of some tutorial action.
In each circuit, the student can measure resistance,
voltage or current, or it can be an erroneous circuit. At this
moment, a circuit is taken as erroneous if the multimeter
setup and the circuit connections are not coherent or if there
is no multimeter. Attending to this classification, we can
visualize the activity of the students by plotting each of the
measured circuits in a timeline (Fig. 9, each column is a
different student). We can see that each student creates a
significant proportion of wrong circuits during the entire lab
time.
Fig. 9. Summary of the actions of each student (columns)
Fig. 10. Creation of one circuit by the students.
It is also interesting to analyze the use of the VISIR
attending to the specific circuits, i.e. for the circuits that are
especially relevant to the academic task.
In the Circuits-based Analysis section of VISIR-DB, we
can select a circuit and see how many times this normalized
circuit was made by each student (Fig. 10). In our study case,
the circuit corresponding to measuring the resistance of a
1-k resistor was built by all the students but students 1 and
17. Student 9 implemented it 18 times.
Although not shown here, VISIR-DB also lists the
normalized circuits by number of implementations. This
allows the instructor to explore whether the expected circuits
have been built. Unexpected and common erroneous circuits
will also show up in this visualization.
As explained above, in the User-specific Results section
of VISIR-DB, we can also select specific users to examine
their work and to provide them more detailed feedback on
their performance and their mistakes.
2019 5th Experiment@ International Conference (exp.at’19)
June 12th – 14th, 2019, University of Madeira, Funchal, Madeira, Portugal
45
Authorized licensed use limited to: b-on: Instituto Politecnico do Porto. Downloaded on February 08,2021 at 16:41:59 UTC from IEEE Xplore. Restrictions apply.
IV. CONCLUSIONS AND FUTURE WORK
As presented in the previous section, using the VISIR-
DB, a teacher can have a better knowledge of what the
students are doing during the learning process. Gained
insights go from time spent in doing an academic task, to
identify students that could benefit from some individual
support, or to identify which specific tasks (expected
circuits) where easily solved and which could need an extra
explanation in class.
In our case study, time on tasks distribution was what we
expected, and some highly-motivated students were found.
Also, a large amount of wrong circuits appeared in the
analysis and this suggests some research may be needed to
deepen into them.
From now on, different teachers that are using the VISIR
remote lab should start to process their own data and then
share these results within the VISIR community [6]. This
process will suggest new functions that can be included into
the VISIR-DB to analyze other aspects of the use of VISIR.
A deeper exploration into the errors and a classification
of the kinds of errors the students are making can also be
warranted by our results. This analysis is likely to modify
how we use VISIR to improve the learning outcomes
attained by students.
ETHICS
Data processing follows the recommendations of the
Ethical Committee of the University of Deusto.
REFERENCES
[1] Gustavsson, I., Nilsson, K., Zackrisson, J., Garcia-Zubia, J.,
Hernandez-Jayo, U., Nafalski, A., Nedic, Z., Gol, O., Machotka, J.,
Pettersson, M.I. and Lago, T. (2009) On Objectives of Instructional
Laboratories, Individual Assessment, and Use of Collaborative
Remote Laboratories. IEEE Trans. On Learning Technologies, IEEE
TLT, Vol. 2, 4, pp: 263-274.
[2] Garcia-Zubia, J., Cuadros, J., Romero, S., Hernandez-Jayo, U.,
Orduña, P., Guenaga, M., Gonzalez-Sabate, L. and Gustavsson, I.
(2017) Empirical Analysis of the Use of the VISIR Remote Lab in
Teaching Analog Electronics, IEEE Trans on Education, IEEE- ToE,
Vol. 60, 2, pp: 149-156.
[3] Marques, M.A., Viegas, M.C., Costa-Lobo, M.C., Fidalgo, A.V.,
Alves, G.R., Rocha, J.S. and Gustavsson, I. (2014). How remote labs
impact on course outcomes: Various practices using VISIR. IEEE
Trans on Education, IEEE- ToE, Vol. 57, 3, pp.151-159.
[4] Serrano, V., Cuadros, J., Garcia-Zubia, J., Hernández-Jayo, U. and
Mompó, L. (2018). Design and Development of a Dashboard for the
Visualization and Assessment of Students Work in a Remote Lab.
IEEE VIS18 (Berlin, 21-26 Oct.)
[5] Chang, W. et al. (2018). Shiny: Web Application Framework for R.
https://CRAN.R-project.org/package=shiny.
[6] International Online Laboratory Awards, Global Online Laboratory
Consortium (nd). Special Interest Group (SIG). VISIR. http://online-
engineering.org/SIG_visir.php
2019 5th Experiment@ International Conference (exp.at’19)
June 12th – 14th, 2019, University of Madeira, Funchal, Madeira, Portugal
46
Authorized licensed use limited to: b-on: Instituto Politecnico do Porto. Downloaded on February 08,2021 at 16:41:59 UTC from IEEE Xplore. Restrictions apply.
