August 9 2018
UNIVERSITY OF AMSTERDAM, FACULTY OF SCIENCE
Thesis Master Information Studies, track Human Centered Multimedia
Visualizing Cat GPS Data
Determining requirements for a visualization of GPS data retrieved from a cat-friendly collar
Sanne Swagerman (11637870)
sanne.swagerman@student.uva.nl
First examiner
Dr. Clara Mancini, Open University Milton Keynes
Second examiner
Visualizing Cat GPS Data
Determining requirements for a visualization of GPS data retrieved from a cat-friendly collar
Sanne Swagerman
University of Amsterdam Amsterdam, The Netherlands sanne.swagerman@student.uva.nl
ABSTRACT
Domestic cats are usually allowed to roam freely outside with-out human supervision, which creates a new market for con-sumer GPS tracking devices. However, current GPS devices and the supporting interfaces and visualizations are usually created with a focus on dog owners. To acquire more insight into the needs of cat owners in relation to GPS devices for do-mestic cats, an iterative prototype based study was conducted. This paper presents the requirements for the visualization of GPS data of domestic cats that were determined in collabora-tion with cat owners. Interviews and multiple prototypes were used as tools to identify and refine these requirements and were mostly based on cat owners’ concerns and curiosity for their cats location. The most important requirements for a cat owner centered GPS visualization were being able to request the live location and being alerted if the cat might be in trouble.
KEYWORDS
human-cat interaction, visualization, gps, geospatial data, user requirements, interface, explorative
1
INTRODUCTION
Most domestic cats are not seen as ”just a cat”, but as a true member of the family, a friend or even an extension of oneself [12]. Pet owners provide their animals with similar levels of comfort, support and affection to that of other family members [10].
Tracking the movements of animals and measuring their vital parameters from a remote location, through bioteleme-try tags or devices worn by animals, is used for studies and individual use. For example, scientists use tracking devices to discover how animals move through their habitat [1][3]. On the individual and commercial side are dog owners who use GPS trackers to keep track of their dogs location for safety reasons [9].
However, a large amount of existing tracking devices are not created with the animal in mind. This can lead to discomfort for the animal which results to altered and unreliable data as well as an impact to their wellbeing [13][14]. That is why Paci et al.[13] at the Open University Animal Computer Interaction (ACI) Lab1created their own GPS tracking collar for cats.
The collar that was designed at the ACI Lab is created with wearability as a number one priority. This means the com-ponents inside the collar are small. To ensure data is being captured for at least 8 hours, the optimal interval of the GPS
1
ACI Lab: http://www.open.ac.uk/blogs/ACI/, accessed: 01-03-2018
device for battery life to log and transmit the data is around once every 30 to 60 minutes. The collar captures and sends coordinates using a sim-card and a GPS component.
The transmitted data needs to be visualized for the owners. Mancini et al.[9] conducted a study with dog-owners to explore the reasons why they were using a GPS device on their dogs. Dog owners may use their GPS device for different reasons and in a different way than cat owners.
A study focused on reasons for cat owners to use a GPS device has not been conducted. Research has shown that there is a significant difference in so-called ”dog people” and ”cat people” [6][7]. This could mean that cat owners have different motivations to use a GPS device than dog owners. Furthermore, dog owners tend to walk with their dogs, while cat owners may use a cat flap for their cats to let them freely roam out-side. Because of this fundamental difference, different user requirements exist.
This paper offers requirements for a GPS visualization for domestic cats, determined in collaboration with cat owners through interviews and an iterative process which used proto-types as a tool. These protoproto-types explored functionalities such as informing the cat owner about the location of the cat and creating zones that indicate a potentially dangerous area or a safezone. Additionally, the prototypes were designed with a system in mind that makes inferences about the wellbeing of the cat so the owners can let their cats out without worrying.
2
BACKGROUND
2.1
GPS devices for cats
Paci et al.[14] conducted a comparative research on GPS de-vices for cats. In this research, they compared two different devices: thePawtrax device, and the Tractive device. They researched if there was an increase in scratching behavior be-tween cats wearing the Pawtrax device, the Tractive device and cats wearing no device. Statistically no significant difference was found, but in some cases, the cats’ scratching behavior changed in a qualitative meaningful way that indicated the cats were experiencing discomfort with the devices [14].
For BBC Two’s Horizon program called ”The Secret Life of Cats”2, a study was conducted on 50 domestic cats living in Surrey. This is one of the biggest studies on the behavior of domestic cats and previously was only conducted on big African cats3. They tracked the cats for multiple days using GPS devices equipped with a miniature camera. They discovered
2Secret Life of Cats: http://www.bbc.com/news/science- environment- 22567526 3
domestic cats share their territory based on a time frame to avoid conflicts. Their results show that often the cats were exploring different areas during their time outside than their owners initially thought they were.
2.2
Existing GPS devices for pets
No scientific research has been conducted with regards to con-sumer use and needs of a visualization of GPS tracking devices focused on pets. Nevertheless, a wide range of pet GPS track-ing devices are currently available on the market. Four of these trackers are Gibi4, Pod35, Tractive6and Whistle7. The main functionality of these devices are locating a pet rather than monitoring activity or heart rate. Table 1 illustrates a compari-son of the visualization part of the data of these trackers.
Comparative
Characteristic Gibi Pod3 Tractive Whistle
Safezone Ë Ë Ë Ë Realtime on demand Ë Ë Ë Ë Directions to location Ë é é Ë Record route Ë Ë Ë Ë Inferences é é é é Automatic check é é é é
Table 1: Comparison of four pet GPS trackers. This comparison shows all of the trackers allow the user to create safezones, see real-time information on demand and record the history of the animal. None of the GPS trackers and visualizations, give the user information about the animals possible wellbeing. All the trackers require human effort to see the location of the animal. The applications do not alert the users when an animal might be in danger. Instead, users have to manually check and make inferences from the data.
2.3
Reasons to track pets
Cat owners consider their cats friends and family [12]. Owners want to know how and what the cats are doing [15].
Reasons to track pets are diverse. For example, a study con-ducted with dog owners by Mancini et al.[9] uncovered several underlying motivations. Some participants of the study ex-plained they tracked their dogs for protection. Protection can mean different things. Participants living in urban areas were concerned about traffc accidents, while participants living in rural areas were concerned about their dogs getting shot. Addi-tionally, regardless of their location multiple participants were concerned about abduction of their dogs. This was especially the case with pure-bred dogs. Another reason for tracking dogs they found in the study was being able to retrieve the dog if they went missing, both on a walk or in familiar surroundings. However, safety concerns are not the only reasons for these dog owners to use a GPS device. Curiosity towards the patterns
4 Gibi: https://getgibi.com 5 Pod3: https://www.podtrackers.com 6Tractive: https://tractive.com 7 Whistle: https://www.whistle.com
of a dog was also stated as a motivation to track a dogs’ location [9].
2.4
Visualizing geospatial data
Geospatial data, is data that refers to a location on Earth, such as a location logged by a GPS device. Maps are commonly used to visualize geospatial data and help users understand this data and relationships within. Using a map, distances, area sizes and directions can be illustrated and patterns can be visualized [8]. Users expect the navigation of a geospatial representation to support a set of basic tasks. These tasks include changing the map projection, its scale, its level of generalization and the field of view by moving, panning and browsing over the map. [5].
Visualizations of geospatial data for consumer use exist in a large number of applications that are designed for running or cycling [2]. Applications, such as Runkeeper or Runtastic (Figure 1)8 9, are based on location and therefore use a map to represent the data.
Figure 1: Left: Runkeeper application before starting a run. User sees the map and knows its location Right: Runtastic application history. User can see the history of the routes he or she has taken, as well as the speed.
3
PROBLEM DEFINITION
3.1
Problem definition
GPS devices that track a pet’s location are increasingly of interest to ACI researchers. The ACI lab at the Open University in Milton Keynes created such a GPS collar specifically for cats. As discussed in section 1, this is a unique device due to its wearability for the cat. However, no research has been done on the visualization requirements and needs of cat owners with regard to the use of GPS tracking devices for their cats.
The goal of this research is to identify a set of character-istics for the visualization of GPS data of domestic cats for consumer use. With this information the ACI lab aims to pro-vide consumers with a user experience that is meaningful and relevant.
8
Runkeeper image: https://support.runkeeper.com/hc/en- us/article attachments/115002089703/unnamed 32 .jpg
9
Runtastic image: https://www.uplabs.com/posts/runtastic- pro- gps- map
3.2
Research questions
As discussed in section 3.1, the ACI lab at the Open University is interested in a user centered research that identifies user motivations and expectations regarding GPS data visualiza-tion for cat owners. The research quesvisualiza-tion identified from this challenge is:How can data, captured by a GPS device at-tached to a cat, be effectively shown to cat owners? By answering this question, this research aims to lay the foun-dation for research that specifically focuses on GPS visualiza-tion for cat owners.
Subquestions that contribute to the research question are: (1) What motivates users to use a system that visualizes
their cats’ GPS data?
(2) What information is relevant to the user?
(3) What are the user expectations with regard to a visu-alization of cat GPS data?
(4) How can a system be created that does not appear to be restrained by a limited battery capacity?
These subquestions contribute to the research question by identifying the motivations and expectations of cat owners in relation so a GPS visualization for cats.
3.3
Relevance
As previously mentioned, no research has been done on the subject of visualizing cat GPS data to cat owners. This research contributes to the knowledge about cat owners and their mo-tivations to know the location of their cats. This research is also an addition to the field of Animal Computer Interaction by revealing user requirements for the system. This informa-tion allows ACI researchers to innovate their GPS device such that it accommodates both user motivations and expectations within a limited battery capacity. As mentioned in section 3.1, research on this particular group of people has not been conducted before. Thus, this subject is a starting point for further research. Additionally, this research contributes to the improvement of user experience for systems related to visu-alizing pet GPS data, specifically that of cats. This is of great value for cat owners who want to use a GPS device for their cats.
4
METHODS
As discussed in section 3, this research aims to explore user motivations and expectations on the subject of cat GPS data visualization. Because of its data-driven nature, this study calls for a qualitative inductive research approach. Furthermore, this research focuses on the case of cat owners, therefore making use of the research strategy of a case study design [4]. To gather the case-specific data, interviews and user tests were carried out.
4.1
Participants
Participants were sampled through convenience due to time constraints and accessibility. All interview participants take care of one or more cats and are currently a resident of ei-ther the United Kingdom (N=5) or the Neei-therlands (N=5). For
the prototype tests, new participants (N=6) and participants who had already been involved in the study contributed (N=3). Three participants for the paper prototype test were completely new to the study. For the second prototype test, two pants had also participated in the interviews and one partici-pant was new to the study. For the third prototype test, two participants had also participated in earlier prototype tests and one was completely new to the study. Participants for the prototype tests also all take care of one or more cats and live in the Netherlands.
4.2
Interviews
Preliminary information regarding cat owners, their cat(s) and their motivations and expectations with respect to a GPS de-vice for cats was collected by means of ten semi-structured interviews.
The semi-structured nature of the interviews gave partici-pants the possibility to elaborate on important topics. There-fore offering insights that reflect what is relevant and important from their point of view. Interviews were conducted in a closed room or a quiet work environment to ensure minimal envi-ronmental disruptions. To properly inform participants on the nature of the research and data processing, they were asked to sign an informed consent form.
The first set of questions served as exploratory questions regarding the cat owner’s lifestyle. Hereafter followed ques-tions related to the cat(s) and their permission opques-tions to freely explore the neighborhood. A third set of questions was con-structed around the imaginary case of using a system that visualizes cat GPS data. These questions focused on users’ pre-liminary motivations and expectations for use. To gain user opinions and ideas with regard to the existing collar prototype, a rundown of the specifications of the prototype was given with a focus on the limitations of the battery life. This was done before asking the last set of questions about a theoretical prototype.
4.3
User testing
To further specify requirements from potential users, proto-types were created and user tests were conducted. The test participants were asked to act as if the prototypes were actual working applications and to think out loud during every test, so everything they thought could be recorded. It was also made clear that there are no right or wrong answers, and the goal was to find for ways to improve the design. For each test, a test plan was constructed containing the required tools to bring, the goals, the introduction and the questions and scenarios.
4.3.1 First prototype. After the interview requirements were set, a paper prototype was created based on these requirements. To test if the application meets the expectations of potential users, this prototype was tested.
This test used the paper on which the prototype was drawn, a laptop for recording responses from the test participants, a printout of the questions and scenarios, and a mobile phone to record the audio of the test. The participants were asked to act
as if the paper was an actual phone, so they could pretend to tap and scroll in the application.
The test began with two questions about the home screen: Q1: What do you think the colored circles are?
Q2: How do you think you can view some more information about a cat on this screen?
After the questions four scenarios were presented to the user: S1: Your cats are outside and you want to see who is where. How would you do this?
S2: You see you have misspelled the name of one of your cats. How do you think you could change this?
S3: You want to know what routes one of your cats takes. How would you do this?
S4: You want to mark a certain location as dangerous. How would you do this?
4.3.2 Second prototype. The second prototype and first dig-ital user test was conducted with the online tool Invision10. This test used a mobile smartphone which could display the Invision prototype, a laptop for recording responses from the test participants and a printout of the scenarios.
For this test the following scenarios were given to the user: S1: You want to know how much battery is left on the tracker of one of your cats. How do you think you can view this? S2: You see you have misspelled the name of one of your cats. How do you think you could change this?
S3: You have the feeling something is wrong with your cat and want to find out where he is right now. How would you do this?
S4: You want to mark a certain location as dangerous. How would you do this?
S5: You get the following notification (your cat entered a dan-gerous area). How do you feel about the text? What would you want to do next?
S6: You get the following notification (cat has not moved for 60 minutes). How do you feel about the text? What would you want to do next?
4.3.3 Third prototype. The third prototype and second dig-ital user test, was also conducted using Invision. For this test also a mobile smartphone with the prototype, a laptop for recording responses and a printout of the scenarios were used. For this test the following scenarios were given to the user: S1: You want to know how much battery there is left on Philips tracker. How do you think you can view this?
S2: You want to view the map without seeing the zones you’ve set. How do you think you can do this?
S3: You want to mark a certain location as dangerous. How would you do this?
S4: You want to go out and do groceries, but you don’t feel comfortable leaving your cats out when you’re away, unless they are in the garden. You will only be gone for a little while, so you decide to view the live location of your cats to see if
10
Invision: https://www.invisionapp.com
they are around. How do you think you can do this? S5: You have the feeling something is wrong with Philip and want to find out where he is right now. How would you do this?
Follow up question: Where do you think the icon stands for?
4.4
Data Analysis
Data from the interviews and user tests was extracted by means of transcription. Hereafter, the transcripts were analyzed us-ing ATLAS.ti11to uncover any emerging themes. All the transcripts were imported in ATLAS.ti and manually analyzed. Codes were created for responses to an interview question or user test scenario. After all the interviews or tests were coded, the amount of times a code appeared was shown by the program. This made participants’ topics and emerging themes easily visible.
5
FINDINGS
As discussed in section 4, one round of interviews and three, iterative rounds of user tests were conducted. This section shows the results and findings. First, the findings of the in-terviews and the requirements that arose are reported. Next, the findings and acquired requirements of the user tests are consecutively shown.
5.1
Interviews
Because of the explorative nature of the study, the interviews focused on the topics of relevant information, device type and compromises regarding battery life. From the interviews two motivations for using a GPS device arose: concerns and curiosity. Some interviewees belong to both groups, some lean more towards one of the two.
5.1.1 Motivations.
”I suppose the main drive would be to see whether their roaming patterns are safe or not.” - Interviewee #3
Owners that largely belong to the concerned owners group were mostly worried about the safety of their cat when they were let outside. What these people have in common is that they all live in cities. Interviewee #1 mentioned that ”there have been some cat related crimes in Milton Keynes and we keep them in until the suspect has moved on somewhere else”. Most interviewees in this group only let their cats out when they are around to check on them. Cars and other vehicles are a big concern for this group. Another precaution taken, as mentioned by interviewee #8, is to let cats wear a light around their neck at night so they are better visible for traffic. Interviewee #9 mentioned she was also concerned about her cat being abducted because she looks like a full breed Norwegian Forest Cat. She also said”If I was living on the ground floor, I would probably not let her outside during the day when I was gone and there was no cat flap in the door. Very selfish.” Even though she thinks it is selfish to keep her cat in, her concern
11
ATLAS.ti: https://atlasti.com/
for the cat’s safety is greater. Some cats were also not allowed to go outside without being on a leash due to safety concerns.
”I am curious where he goes, I think that it’s really nice to see. Just out of curiosity.” - Interviewee #8
Owners that largely belong to the curious owners group live in both rural and urban areas. They are mostly interested in what their cat does outside of the house and are not as concerned for their safety. They think they have a general idea of where their cat goes but would very much like to know where they go exactly.
5.1.2 Worrying. Some participants state it might ease their mind to know where their cats are. The reason interviewee #9 would use a GPS device is”to give me more of an ease of mind, to not stress over her safety or where she would be”.
Other participants found it may actually achieve the oppo-site effect. Interviewee #5 explained”I think that a GPS device can increase your anxiety. Once you have the tool, you know you can check, you are looking more often and you see all kind of things”.
5.1.3 Device. All participants explained they are in favor of the application being on a mobile device, such as a smartphone. Participants explained the reason for this is due to the fact that they have their smartphone almost always with them. This way, if something happened they can not only get notified immediately, but can also go retrieve the cat.
5.1.4 Notifications. The idea of receiving notifications about a cat not moving or entering a set location, came forward as an important functionality for the interviewees. But none gave the exact same threshold for when to receive these notifica-tions. Some said they would want to receive a notification if their cat has not moved for ten minutes, others said they would want to receive a notification only if the cat has not moved for more than three hours. Interviewees also mentioned they want the ability to set the threshold themselves, to play around and figure out what a good interval is.
Another area of variety in responses regarding notifications, is when the cat goes in or out of the house. Some interviewees mentioned this would ease their mind and be helpful for them, others said this would be perceived as annoying and too much information. Again, the interviewees mentioned they should be able to change the frequency of this notification themselves. 5.1.5 Route. Interviewees mentioned they would like to see the route their cat has taken on that current day. They are also interested in the route history.
5.1.6 Live tracking. The subject of live tracking is one that appeals to many interviewees. Mostly from the perspective of safety, sometimes also due to curiosity. All participants mentioned live tracking their cat would be optimal.
5.1.7 Requirements. With these results from the interviews, the following requirements were set for the application:
• Live tracking
• Alert when cat goes into and out of a dangerous place • Alert when cat goes into and out of the house • Be able to customize the notification settings • Notify when cat might be in danger • Route cat current day
• Route history
5.2
Paper prototype
The goal of this test was to find out if the requirements that came out of the interview were met and if there were functions missing in the prototype. All of the scenarios were created with the purpose of letting the user navigate through the application to find out if functions were logically placed.
5.2.1 Test results. The idea that the colored circles (Figure 2) are the cats was clear. Participants explained this was mostly due to the circles being on a map.
Figure 2: Left: Map view of the cats. Right: What hap-pens when a user clicks on one of the cats
When asked to get information about the cat, every par-ticipant clicked on the circle. When they wanted to get more information, every participant tried clicking on the tooltip that was opened. This felt more intuitive for them than using the paw menu item, which was intended to give them more information (Figure 3).
None of the participants could figure out how they could change the name of a cat. All thought they could do this by clicking on the name of the cat, instead of clicking on the pencil icon.
The placement of the route history of a cat was also in an intuitive place (Figure 4). All participants expected the cat specific route information to be on the detail page of this cat. Participant #3 explained she would like to see exactly at what time the cat was in a certain place. The way you could select a time period for the history of the route of a cat was not optimal. One participant explained he would prefer to see some sort of calendar.
What was clear from the test was that the way to set zones should be different. In the paper prototype the option to mark a place as dangerous was put in the ”Settings” page, but this
Figure 3: Left: Cats overview after user clicked the paw icon. Right: Cat Detail page with cat information. seemed not intuitive to the test participants. Participant #2 explained that the locations you want to set are relevant to the cats, so he would expect this option to be at the ”Cat” page.
Another area that was very clear during the tests was that the used icons were not the best options. Especially the icons for the ”Map” and ”Cat” page were not clear.
All of the participants explained they would like to get a notification when the cat did not move for a set amount of hours.
Figure 4: Left: Cat Detail page scrolled to the bottom. This reveals the View History option. Right: The route history of a cat.
5.2.2 Requirements. The first user test resulted in new re-quirements for an application:
• Navigating to the cat detail page by clicking on the tooltip of a cat
• Select date by using a calendar • See the battery level of the device
• Get a notification if the cat does not move for a certain period
• Use of intuitive menu icons
• Move Set Zones to the Cat Overview page
5.3
First digital prototype
The main goals of this test were to determine if the new place-ment of functions had any influence on the navigation flow in the application and findability of functions.
5.3.1 Test results. Three participants explained they would like to see the battery level of the GPS device on the map view, rather than having to click to the Cat Detail Page (Figure 5).
Figure 5: Left: map view of the cats. Right: Cat Detail Page with information about the cat and the option to request a live location.
Because the edit button in the paper prototype was not identifiable as a pencil, for this prototype it was chosen to keep the pencil and see if with a more recognizable icon the function is clear. When the participants were asked to change the name of a cat, the edit button was found this time, even though the positioning had not changed relative to the paper prototype (Figure 5).
The introduction of the live location functionality was well received. Although the real time tracking of the cat was really liked by all the participants, the positioning was a bit confusing (Figure 5). Participant #5 pointed out that if she was really stressed, it would take too many steps to get to the desired function. Also, the icon that represented the functionality to open the location in another application, such as Google Maps, was not clear (Figure 6). Participant #5 said”I did not understand what the car meant, I thought it might mean he is near a road”. Another comment on this functionality, was the option to request the live location of all of your cats at one time, instead of requesting them one by one. Participant #6 mentioned she would like to be able to share the location of her cat with friends, so they can help find the cat.
Even though the positioning of the ”Set Zones” function had been moved to the ”Cat Page”, test participants could not easily find it. They rather searched for it on the map and settings pages. Participant #6 explained the Cat Overview page was the last place she expected to find it, because the zones are map related (Figure 7).
Figure 6: Left: After a user requests the live location of a cat, the live location page is shown. After the user clicks on the cat, this tiptool is shown and allows the user to get directions to the cat. Right: The route history page of a cat.
The creation of new zones was also not intuitive. It was not clear the participants had to press the OK button and the check icon to confirm the new zone (Figure 7).
Figure 7: From left to right: 1: Overview of all zones. 2: A shape can be chosen for the new zone and the position can be changed by dragging the shape. Releasing the shape triggers a pop-up that lets the user choose a name and whether it is a dangerous area or a neutral area. 3: After the user confirms the new zone, it is visible on the overview page.
All the test participants mentioned they wanted to be redi-rected to the live location of their cat when they receive both the ”Cat has entered dangerous location” and ”Cat has not moved for x amount of time” messages. One participant men-tioned she liked the fact she could already see for how long the cat has not moved from its position, to make the decision herself if it is urgent or not.
5.3.2 Requirements. The first digital prototype lead to new requirements:
• Show GPS battery level in the map view • Be able to share the live location of a cat • Be able to set the zones from the map view • Be able to access the live location of a cat quicker
5.4
Second digital prototype
The main goal of this user test was to verify if the function of setting zones and retrieving the live location of a cat were working more intuitively.
5.4.1 Test results. The complete prototype can be viewed using the following link: https://invis.io/XMKUC9DV2CW
During the last test it became clear that the live location re-quest functions of a cat, setting zones and checking the battery level were best placed on the map itself. All test participants reported it was intuitive to click on the menu item for these map related actions (Figure 8). Test participant #5 who was also present for the first digital prototype test explained”I like that everything that has to do with the map is accessible through the map view and that I don’t have to go to the cat page every time”.
The additional battery percentage on the map view was also received well. Especially because this is quite an essential part of using the device, and now it is quickly visible upon opening the application (Figure 8).
Figure 8: Left: Map overview with cat locations. Ex-panded menu and shows options to toggle zone visibil-ity, edit zones or request the all cats’ locations. Right: Tooltip appears when a user clicks on a cat. Refresh but-ton allows the user to request the live location of this cat for five minutes.
When users are worried about their cat, they can open the application and easily request the live location of one or all of their cats on the map overview (Figure 8). One comment was that instead of the live location being active for five fixed minutes, users might prefer a toggle (Figure 10). While the participant was not opposed the fixed five minutes, she men-tioned it might be a waste of battery in the collar, in situations such as scenario four (4.2.3).
Figure 9: From left to right: 1: Edit zones main page. 2: Creating a new zone. 3: Naming and coloring a zone. 4: Map with newly created zone.
When asked to mark a location as dangerous, the partici-pants all clicked on the menu and navigated to the menu item ”Edit Zones” (Figure 9). By removing the OK button in the tooltip from the previous prototype, it was easier for the par-ticipants to create a new zone. Participant #6 mentioned the process could be simplified by leaving out the verification of a newly assigned position and immediately showing the tooltip (Figure 9, second and third image).
Participant #7 who did not take part in the interview or an earlier user test mentioned it would be good to have a setup stage when the device and application are first used. Namely because the zones were not explained. Even though she was able to find how to edit them and add new ones, she did not fully understand the purpose intuitively.
Figure 10: Live location is active for five minutes and visible from the map. Left: Map view when one cat’s live location was requested. When clicked on this cat the tooltip appears with the option to share the location to another application or person. Right: Map view when all live locations are requested.
The icon for sharing or opening the live location of the cat was also much clearer (Figure 10). When asked what the icon meant, all of the participants immediately mentioned sharing the location to either an application or a person. One
participant mentioned”I know what this icon means because it is also used in other apps for the same action”. The participants also mentioned they knew which cat was currently showing the live location, because they understood what the green dot next to the colored circle indicated.
5.4.2 Requirements. Finally, the last prototype refined the existing requirements:
• Simplifying creating new zones by combining steps • Enable users to determine the duration of the live
location (battery life)
• Create tutorial for using the application, with empha-sis on the zones
6
DISCUSSION & CONCLUSION
6.1
Discussion
The requirements that were determined in this paper can be used for an application for consumer use. This research allows designers and engineers to create a working system for cat owners to use without worrying that the system will affect the cat in a negative way in his or her life.
6.1.1 User motivations. The aim of this study was to iden-tify several requirements for cat owners regarding visualization of GPS data. Analysis of interviewees answers suggest two im-portant reasons to use a GPS system for cat owners are safety and curiosity. The owners are concerned for the safety of their cat, especially in regard to traffic. Additionally, they are cu-rious to find out where their cats roam during the day. This results in functions such as the ability to request live location and receiving a notification if the cat might be in danger. While the reasons for cat owners to use a GPS system are similar to that of for example dog owners, this is an important difference in usage requirements.
6.1.2 User expectations. Cat owners prefer to receive noti-fications about the cat’s location, instead of manually having to check the application throughout the day. They would like to get notifications if there might be something wrong with their cat based on the amount of time they have not moved, while also being able to check the (live) location and history of the cat themselves. It is important that these notifications can be adjusted by the users because cats have different activity patterns while they are out of the house. The adjustment of notification intervals has the potential to be automated by ma-chine learning algorithms. Additionally, because the threshold for the interval of the notifications differs so much from person to person, there is no middle ground that would satisfy all.
To increase the likelihood of a cat being found quickly if something might be wrong, the live location of a cat should have the option to be shared to another application or person. Also, the function to request the live location of a cat should be easily accessible so users do not have to click too many times to get to an important function. This corresponds with studies done on the impact of anxiety on cognitive functions. Research suggests anxiety reduces the ability to focus on anything else than the apparent threat. When the user is unable to find the
function to locate the cat, this could worsen the anxiety and fear [11].
Setting (dangerous or normal) zones should be done on the map. Since these zones are linked to the map, it was logical to the test participants to have the functionality at the same place. The user should also be in charge of the notification options regarding the zones.
It is interesting to think about ways to extend the zones beyond notifications. The collar could for example emit some kind of feedback to the cat to condition him or her to not go to the area. In this case, the ethical implications need to be considered. This addition to the collar would mean it limits the cats’ ability to move freely while being outdoors and has the potential to change their inherent behavior.
6.1.3 GPS collar possibilities. Due to the limited battery capacity of the collar, a trade-off needed to be found between the interest of the animal (an animal friendly collar), and the needs of the users. A function was proposed that initiates a live location of the cat for five minutes. As described in section 5.4.1 and 5.4.2, additional battery life could be saved by giving the user the option to turn live location off within five minutes. If battery life were to increase, an addition to the collar could be a heart rate monitor or other biological data. This would allow the system to make even better inferences about what the cat is doing. For example, if a cat is sleeping the heart rate would be lower than while walking, while the GPS location stays the same. If the heart rate is high, and the GPS location stays the same, the cat could be in distress.
The collar could even be extended in such a way that it enables the cat to be autonomous when the owners are not home. For example, opening the cat flap and activating the GPS when going outside and feeding itself.
6.2
Limitations
A clear limitation of not necessarily the application itself, but the system as a whole, is that the GPS data of a cat is only captured and sent every 30 to 60 minutes. Multiple intervie-wees expressed this would be inconvenient for them, because it would not ease their mind about the safety of the cat. Also the people who would be more interested in a GPS device out of curiosity for the routes a cat would take, expressed this time frame is quite large. In 60 minutes a cat is able to cover a large amount of land, and this may not be accurately presented if the location is only recorded once or twice.
The reason this interval is around 30 to 60 minutes has to do with battery size and capacity. This is not something that can be easily changed in the short term, but it would be ideal if the GPS capturing and sending interval could be made shorter.
6.3
Conclusion
This study explored how data captured by a GPS device, at-tached to a cat, can be shown to the owners of domestic cats. The most important requirements for a visualization for cat GPS data have been identified. To achieve this, the study explored motivations and expectations of cat owners. Two motivations for cat owners for using a GPS visualization based on their cats
are concerns and curiosity. Some people overlap, while others lean strongly towards one motivation. Cat owners expect to receive notifications about the location and possible well being of their cat. Additionally, they would like to be able to set zones and receive notifications when a cat enters one. Furthermore, cat owners want the ability to request the live location of their cat.
With this information, ACI researchers at the Open Univer-sity in Milton Keynes can create a consumer product which is not only cat-friendly, but also provides a good user experience to the users and fits their expectations and needs.
6.4
Future work
Future research is needed to look at ways to adapt the infer-ences that are made about a cat to the individual, so notification intervals do not have to be set manually. Some cats sleep more often and are not that active, other cats may prefer to sleep in a few known spots and are very active. This means the alerts that are send out could be false and lead to unnecessary worry for the owner.
Additionally, the research needs to be repeated on a larger scale to confirm the requirements as identified in this research and to be able to generalize the data. In this case, it would be advisable to use a different type of sampling to understand the habits and requirements of different cat owners, such as prob-ability sampling. By using probprob-ability sampling, all different types of cat owners would get a chance to be able to participate in the research.
ACKNOWLEDGEMENTS
First of all, I would like to thank my supervisor dr. Clara Mancini for her support, feedback and hospitality in Milton Keynes. I learned a lot from our chats about the interviews, the feedback for the structure and the interesting sessions I was able to attend.
I would also like to thank dr. Frank Nack for arranging dr. Mancini’s trip that makes it possible for her to attend the defense of my thesis at the University of Amsterdam. Addi-tionally, I would like to thank him for initiating the contact between me and dr. Mancini.
Next, I want to thank my friends and co-students for their patience and help with the project. I would like to single out Koos Klooster and Maaike Koolbergen for their extensive proofreading of this paper.
Additionally, I would like to thank my parents for their support, proofreading and contributions that made my trips to Milton Keynes possible. While on the subject of family, I would lastly like to thank my dog No¨el for the countless cuddles and stress-relieving play times.
REFERENCES
[1] Aebischer, N. J., Robertson, P. A., and Kenward, R. E. (1993). Compositional Analysis of Habitat Use From Animal Radio-Tracking. 74(5):1313–1325. [2] Bauer, C. (2013). On the ( In- ) Accuracy of GPS Measures of Smartphones : A
Study of Running Tracking Applications.International Conference on Advances in Mobile Computing & Multimedia, (April):335–340.
[3] Bridge, E. S., Thorup, K., Bowlin, M. S., Chilson, P. B., Diehl, R. H., Fl ´eron, R. W., Hartl, P., Kays, R., Kelly, J. F., Robinson, W. D., and Wikelski, M. (2011).
Technology on the Move: Recent and Forthcoming Innovations for Tracking Migratory Birds.BioScience, 61(9):689–698.
[4] Bryman, A. (2012).Social research methods. Oxford university press. [5] Cartwright, W., Crampton, J., Gartner, G., Miller, S., Mitchell, K., and
Siekier-ska, E. (2001). Geospatial Information Visualization User Interfaces Issues. Cartography and Geographic Information Science, 28:45–60.
[6] Gosling, S. D. and Bonnenburg, A. V. (1998). An integrative approach to personality research in anthrozoology: Ratings of six species of pets and their owners.Anthrozoos, 11(3):148–155.
[7] Gosling, S. D., Sandy, C. J., and Potter, J. (2010). Personalities of self-identified ”dog people”and ”cat people”.Anthrozoos, 23(3):213–222.
[8] Kraak, M. J. and Ormeling, F. (2010). Cartography: Visualization of Geospa-tial Data.
[9] Mancini, C., Linden, J. V. D., Bryan, J., and Stuart, A. (2012). Exploring Inter-species Sensemaking : Dog Tracking Semiotics and MultiInter-species Ethnography. UbiComp ’12, pages 143–152.
[10] Martens, P., Enders-Slegers, M. J., and Walker, J. K. (2016). The emotional lives of companion animals: Attachment and subjective claims by owners of cats and dogs.Anthrozoos, 29(1):73–88.
[11] Mathews, A. (1990). Why worry? The cognitive function of anxiety. Behaviour Research and Therapy, 28(6):455–468.
[12] Noz, F. (2011). Cat cat revolution: An interspecies gaming experience. Proceedings of the 2011 annual conference on Human factors in computing systems, pages 2661–2664.
[13] Paci, P., Mancini, C., and Price, B. A. (2016).Towards a wearer-centred framework for animal biotelemetry.
[14] Paci, P., Mancini, C., and Price, B. A. (2017). The role of ethological observa-tion for measuring animal reacobserva-tions to biotelemetry devices.ACM International Conference Proceeding Series, Part F1325.
[15] Yonezawa, K., Takashi, M., and Rekimoto, J. (2009). Cat @ Log : Sensing Device Attachable to Pet Cats for Supporting Human-Pet Interaction.ACE ‘09: Proceedings of the International Conference on Advances in Computer Entern-tainment Technology, pages 149–156.
