Visualizing Data Trails: Metaphors and a Symbolic Language for Interfaces

Visualizing Data Trails: Metaphors and
a Symbolic Language for Interfaces
Omar Sosa-Tzec
University of Michigan
Penny W. Stamps School of
Art and Deisgn
Ann Arbor, MI 48105, USA
omarsosa@umich.edu
Abstract
This paper starts by presenting four mobile interface
design concepts to make personal data trails visible.
Frosted screen, rainbow heatmap, hungry zombie, and
data current are the labels given to these concepts.
After reflecting on these concepts, the paper focuses on
some elements of personal data trails, and explores a
visual system of icons to indicate the user the possible
use and abuse of the data they produce during the UX.
The elements proposed by this paper are surveillance,
commodification, data aggregation, data input, affect
and arousal, preferences, and community.
Author Keywords
Personal Data Trail; Communication Design;
Information Design; Visual System
CSS Concepts
• Human-centered computing~Interaction
design~Interaction design theory, concepts and
paradigms
Speculative Designs for Emergent Personal Data Trails: Signs,
Signals and Signifiers
Workshop at CHI 2020 that took place online on April 22nd
, 2020
(in response to Covid-19)
Website: https://www.emergentdatatrails.com/
Workshop summary on the ACM DL:
https://doi.org/10.1145/3334480.3375173

User Interface Metaphors for Visualizing
Personal Data Trails
My first response to the workshop’s call was to
conceptualize appearance and behavior that help the
user become aware of the overall amount of data they
produce and disperse among the apps they use daily.
Four proposals emerged from this exercise (Fig. 1). The
first design concept centers on the idea of a frosted
screen (Fig. 1A). The degree of frostiness indicates the
amount of data in circulation and used by third parties.
For those times in which the phone activates
automatically and the screen lights up, the degree of
frostiness will warn the user about how much data is
being stored in the phone and deployed over the
Internet for others to use. If the user decides to use
their phone anyway, the frostiness still shows to remind
the user about this situation. The frostiness drastically
changes as the user changes of app. Each app thus
involves a particular degree of frostiness while the lock
and home screen show the aggregated use and abuse
of personal data. Such a change allows the user to note
which app might be compromising their privacy.
The second concept is a rainbow plasma cloud (Fig.
1B). Instead of showing a degree of risk as the frosting
screen concept does, the plasma cloud concept
visualizes several elements that characterize a personal
data trail. One of these elements could be identity. For
this element, the plasma cloud would visualize
variables related to what defines the presence and
behaviors of the user in the Internet. For example, the
plasma cloud would visualize how much data related to
cultural affinity, physical activity, or community
engagement the user has produced as a result of using
different apps during the day, week, or month.
Moreover, the plasma cloud would visualize the degree
of mobility of the user based on the different places
they have been lately. Further, the plasma that
represents mobility would visualize predictions of how
mobile the user might become soon. The plasma cloud
would visualize the degree of commodification applied
to the user as well. This represents how valuable the
user’s data currently is for third parties, which have a
business relationship with the company that created
the app. This plasma would show both the current and
predicted value of the user regarded as a commodity.
The plasma cloud would be also able to visualize the
degree of arousal in the user based on their phone
usage during a certain period. Further, this plasma
would visualize the potential degree of arousal in a near
future.
The rainbow plasmas cloud appears as a dynamic layer
on the screen of the phone (Fig. 1B). When the user
performs an action, the interface components involved
in it will define the area wherein more multicolor
plasma will emerge. Immediately, the plasma will
divide and each portion will join the mass having its
same color. The extent of each color thus indicates the
amount of or significance of a particular data kind, such
as identity, mobility, commodification, and arousal (see
legend with different color swatches on Fig 1.B).

Figure 1: Interface metaphors for visualizing personal data trails: (A) frosted screen, (B) rainbown heatmap, (C) hungry app zombie,
and (D) personal data current.

The third concept that came to my mind is regarding
product (app) as a zombie (Fig. 1C). Each zombie
feeds on data, increasing its size in function of how
much data it eats. The data that works as food is the
data that an app produces. The more data the app
produces, the bigger and chubbier the app’s icon
becomes. As it increases, the icon starts occupying
space on the screen, occluding the other icons. The
visualization deforms each icon to make the user
become aware of the degree of monstrosity that an app
entails when it comes to the production and store of
personal data. This visualization, however, does not
show the data connections between an app and others.
To emphasize these connections, I came up with the
concept of data as current (Fig. 1D). This visualization
still regards each product (app) as a creature,
somehow similar to slug or octopus. The personal data
produced by each app makes the icon of the app
develop tentacles which, by suction, connect with the
icons of the apps that use and benefit from the data.
Once the connection has been established, the user can
notice how the data flows. The data appears as bright
little dots—a visualization that is similar to the one
offered by those USB cables with LEDs within that show
a flow of light dots going from the outlet to the device
to indicate the presence electric current. The frequency
and velocity of dots sent from one app to another
indicates the amount, sensitivity, or significance of the
data being produced and shared between products.
Communicating the different kinds of
personal data trails through symbols
After completing the ideation exercise, I reflected upon
how these design concepts can help me give a response
that is closer to the workshop call, which I understand
it is more interested in concrete signifiers to indicate
aspects of personal data trails. The first thing I noted is
that personal data trails need to be characterized. In
my first exercise, I focused not only on the production
but also the use and abuse of user data by the different
products that a user has on their phone. During and
after completing the exercise above, I realized that, for
me, an approach to designing signifiers to raise
awareness in users requires the definition of elements
that (might) constitute a personal data trail.
As I am concerned with the production, use, and abuse
of personal data, I considered for my next exercise to
focus on how the use of a product entails a risk for
the user based on the kind of personal data trail
the use generates. Hence, I propose to consider two
states of use: (a) when the user is not using the
product (at least not consciously), and (b) when the
user is (consciously) using the product. I considered
that a user would easily recognize the level of risk
entailed by the generation, use, and abuse of their data
by referring to three simple levels of risk: (a) null or
low, (b) moderate, and (c) high. The null/low level
implies that a product produces and stores temporary
data. The use of this data is quite local and is likely to
be irrelevant for other products. The moderate level
indicates that a noteworthy personal data trails is or
will be generated as a consequence of using the
product or one of its features. The data generated as a
result of using the product implies a potential negative
effect on the user’s integrity—an exposition and
manipulation on who the user is and can become, what
the user does and wants, why the user behaves in a
certain way. This level indicates that the user’s data
trail can benefit other products, and third parties in
general. If the data produced and shared can have a

direct effect on the user’s integrity, then one talks
about a high level of risk. The data produced, stored,
and shared by the product put together, can be useful
to influence in the user’s political perceptions, good
consumption, social benefits, personal aspirations,
mood, and other personal and sociocultural aspects
that define the user as an individual of both the real
world and the Internet.
Based on this classification, I created different signifiers
to communicate the relation between the risk entailed
by the production, use, and abuse of personal data,
and whether the this situation involves a conscious use
of the product (Fig. 2).
Figure 2: Representation of the risk as the density of a user’s
personal data trail, which may be a result of actively using the
product or not (inactive).
The density of the trail depicted in these signifiers
makes reference to the idea of data current introduced
above (Fig. 1D). Certainly, the product may not
generate too much data and yet there is a chance that
the use and abuse of such data can affect the user
considerable. For simplification purposes, I use the
density of the trail to indicate that regardless of the
amount of data, data can flow quickly beyond the
user’s awareness.
Mostly inspired in the rainbow plasma cloud (Fig. 1B)
and how it aims at visualizing different elements of a
personal data trail—its potential use and abuse; what
to do with it. I define a classification, including its
signifiers, for the use and abuse of the data constituting
a trail (Fig. 3). This classification comprises
surveillance, commodification, data aggregation, data
input, affect and arousal, preferences, and community.
Figure 3: Classification of the potential use and abuse of data
constituting a personal data trail.

The signifier of surveillance indicates that the product
and its business partners purposefully enquire into the
data produced, stored, and shared as a result of using
the product. The signifier of commodification indicates
the user that the product and its business partners
obtain a financial benefit from such data. The signifier
of data aggregation is for the user to know that all the
data produced will be aggregated to larger data bases.
This category suggests that a certain level of anonymity
is involved in the process of sharing and using the data
by the product and its business partners. The signifier
of data input is to tell the user that all the data
produced, stored, and shared is used in machine
learning algorithms, not necessarily related to the
product’s functionality. In other words, this category is
to remind the user that the data is being used to feed
algorithms whose aim is to discover a pattern that
benefits the company of the product and its business
partners.
The signifier of affect and arousal indicates the user
that the data produced, stored, and shared can help in
determining its current and expected emotional state,
and that knowing such state is beneficial for the
objectives of the product and its business partners. The
little star of the signifier for preferences reminds the
user that all the data produced, stored, and shared can
affect content and consumption preferences. Finally,
the signifier of community tells the user that their data
is used to identify, characterize, or even define who
they are as an individual in both the real world and the
Internet. This category is concerned with how the user
data is used to position them in certain demographics,
which in consequence, will involve a particular set of
opportunities and limitations at the personal and social
levels.
With this visual elements defined, I created and
propose a visual system to indicate whether the
personal data trail is a consequence of a conscious use
of the app, the level of risk that the production, use,
and share of data entails, and the use and potential
abuse of the data (Fig. 4). This system provides a
series of interface icons that a designer can be used to
Figure 4: Subset of the visual system proposed. This subset shows the combination of the two signifiers indicating a high level of
data risk with the signifiers showing the potential use and abuse for the data produced by a trail.

demonstrate an ethical commitment to safeguarding
the integrity of the user and the data they product
while they use the product. The designer can integrate
these icons to the section of terms and condition. The
user can also integrate them in pop-up windows or
messages from time to time to remind the user about
the implications of using the product.
Discussion
In this paper, I showed my response to the workshop
call, which is twofold. The first part of my response
involved the ideation of some design concepts in which
personal data trails can become visible and disruptive. I
would say that these concepts are motivated by need of
making people aware of the production, use, and abuse
of the data derived from interactions with apps and
other forms of technology. We cannot judge what we
cannot see. These four metaphorical implementation
seek to delightful and critical at the same time.
One point that I want to emphasize is the use of
sketching as an ideation tool/skill. I consider that for us
to come up with interesting signifiers, we need proper
ideation tools and skills. Starting my exploration with
sketching allowed me to be imaginative without the
pressure of coming up with a right solution; especially,
because I am not sure if my understanding of personal
data trails is the same the facilitators of the workshop
have. However, sketching allows to produce material to
discuss ideas and, in this particular case, to reflect
about the constituents of a personal data trail.
Once I concluded my sketches, I felt I have my eureka
moment when, from my perspective, I found that can
articulate the ideas from these four design concepts as
a list of elements and characteristics of a personal data
trail. Sketching was again a significant activity.
However, in this case, I carried out my sketching
directly on the computer because I considered
important to think how the interface icons would look
like as part of a user interface. During the whole
process I had in my mind the creative commons icons
as the visual reference—based on one email sent by
one of the facilitators to the PhD email list. I think that
having such a reference was important for this second
part of my response. My aim was not only finding an
appropriate signifier but also shaping it in such a way it
appears clear and neat, so it can look serious,
somehow authoritative, and universal.
My main takeaway is that, in order to generate a visual
system of signifiers, we need to identify or define what
we want the user to be aware of, and later, to identify
universal symbols that many users can understand.
Another interesting thing that guided my design intent
is frame the visual system through a subtle lens of
criticality, not to say fatality. I found interesting and
relevant that designers involved in the creation of this
kind of signifiers raise awareness and emphasize how
delicate and harmful the abuse of personal data trails
can be.

Visualizing Data Trails: Metaphors and a Symbolic Language for Interfaces

More Related Content

Similar to Visualizing Data Trails: Metaphors and a Symbolic Language for Interfaces (20)

More from Omar Sosa-Tzec (20)

Recently uploaded (20)

Visualizing Data Trails: Metaphors and a Symbolic Language for Interfaces