SJournal

02830 Advanced Project in Digital Media EngineeringSJournal - a part of SensibleDTU

Ulrik Uhre Brinks081830

[email protected]

Kristian Michael Clarksons072660

[email protected]

ABSTRACTThe main problem that our application will address is, that itcan be difficult to remember certain events; such as where theuser was last week and who the user talked to last night atwork. This project will search to solve the problems concern-ing remembering places and interactions with other people.

Our application will try to solve the problem while allowingthe user to go through the data by giving them a simple andintuitive interface and only present the most relevant data.

Combining gathered data and a great looking interface, wewill be able to give the students at DTU a prototype of a mo-bile application, that will make it easier to gain insight abouttheir places and interactions.

Author KeywordsInterface; Quantified Self; Android; User Centered Interface(UI); User Experience (UX); Lifelogging

INTRODUCTIONThe prototype application, SJournal, is part of the Sensi-bleDTU program and in this project try to come up with aimproved version of the existing Android application Sensi-bleJournal[4].

Our vision for improving on the existing application is to cre-ate a whole new way of presenting the same gathered data. Inorder to improve it, we will look at the market and see howsimilar applications are presenting their data. The prototypeapplication will allow the users get an easy overview of theircollected data and allowing them to gain more insight of thedata. Since only students at DTU are allowed access to thedata, it would be interesting to allow them to see who theyinteract with and in what context they interact. SJournal canbe described in a tweet (140 characters) which is illustratedin Figure 1.

Our motivation for doing this project is our deep curiosityabout gaining self knowledge of our activities through datacollected from different sensors. The technology behind lifel-ogging has evolved the past year and lifelogging is getting

Figure 1. Vision of SJournal described in only 140 characters

more and more interesting. Cultural trends shows that log-ging your every move is getting more socially acceptable[8]and with devices like the JawBone Up[15], FitBit[16] andmobile tracking applications like Moves, it gets easier to trackyourself and therefore even more popular[9]. Life logging isabout digging in to the data gathered and getting an insight ofwhat and when certain events occurred.

Besides our motivation for data and self tracking, we havebeen working on a project in collaboration with GlostrupHospital concerning people diagnosed with MCI[7]. Thesepeople are having trouble remembering daily tasks, and lifel-ogging would in a near future be an assistance to them. Dur-ing interviews they have specifically asked for a way of log-ging their daily routines and help them remember earlier oc-curred events.

As we are working under a restricted time frame we will notbe able to both design and implement a vertical prototype[11].We have chosen to go in depth with, what we believe is themost important aspect, the user interface. To get a feeling ofwhat kind of data is available from the SensibleDTU we willalso develop a prototype application including a JSON-parser.This will allow us to gather the most relevant data and test ifit is feasible to solve the problem concerning rememberingplaces and interactions with other people.

The report is build up around the prototype including relatedwork which includes a market analysis on similar applica-tions, methods on how we did the project, results from ourwork, discussion and evaluation on our results and a con-clusion including future development.

1

RELATED WORKBefore beginning the development of a new application andto improve the existing SensibleJournal application, it is im-portant to gain knowledge on the existing market of similarapplications to see how they do it.

The following section will include an analysis of the currentavailable applications and an evaluation on their strengths andmain focus. This evaluation makes it possible to create anoverview of what our application should include in order toimprove on the existing SensibleJournal-application[4].

Related apps on the Android-marketThere are currently several different markets for mobile ap-plications under a variety of mobile platforms including iOS(iPhone), Android and Windows Phone. In this project wewill be focusing on the Android-market, as this is our pre-ferred markup language.

It is important to have a good knowledge of any competitionwhen a new application is developed, which should take thebest and differentiate from the rest. The following is there-fore a review of the applications which are considered as thelargest competitors in terms of the new application.

Moves [1] allows the user to record and track their current ac-tivity i.e. walking, running, cycling and transporting. Movescan create a storyline for each day including the places theuser has been and what type of activity. The main focus istracking the activities between places. An example of the sto-ryline and summary of activities is illustrated in Figure 2.

Figure 2. The daily storyline and maps show where, when, and how youmove).

Friday [2] allows the user to get a complete log of i.e. placesvisited, people interacted with, music listened to, calls, emailsand texts. All of these data is logged and will help remember-ing past activities, people talked to and events attended. Themain focus is to tracks all possible activities in a lifeloggingjournal. The timeline of interactions is illustrated in Figure 3.

Trails [3] is created by the same developers who also de-veloped the Friday-app and allows the user to automatically

Figure 3. The timeline of people interacted with during a given day.

mark life events based on specific locations. In a map-viewthe user is able to get an overview of each day. The main fo-cus is to track the users whereabouts without disturbing theuser. The user is able to extend life-logging with data fromi.e. social networks and calls. The timeline of interactions ona given place is illustrated in Figure 4.

Figure 4. The timeline of people interacted with on a given place

SensibleJournal [4] allows the student at DTU to get anoverview of visited places each day and how long they havebeen there. Furthermore, it includes the time used on differ-ent activities i.e. stationary, walking and ”on a vehicle”. Thetimeline is illustrated in Figure 5.

Summary of related workBased on our small market analysis of the existing apps wehave become aware of what applications we would stand up

2

Figure 5. The timeline of time spent at places and on activities

against if releasing the application. The following are the fea-tures that our application should include as the absolute min-imum in order to compete on the market with self knowledgeapplications.

• a simple and intuitive interface

• interactions with other users to some extent

• a history of places visited and people interacted with

METHODSOur workload for developing the prototype is divided intoseveral smaller parts. The following section will include theuser-stories, use cases and the steps of developing the proto-type.

User storiesThe users are a crucial focus area when developing a user-centered application. Our core users needs and values shouldbe taken in consideration when designing the interface andadding features to it, as they would be the end users to theapplication.

To be able to specify our users, we are making use of user sto-ries and use cases, and based on these we will make sure thatthe included features in the prototype application will meettheir needs and requirements:

• As a student I can quickly check my timeline, so I can bereminded what happened and who I have talked to at theparty last night.

• As an elderly person I can see who I have talked to today,since I tend to forget.

• As a life loggist[5] I can go back in time, so I can see earliervisited placed and check which places I visit the most.

• As a consultant I can check newly visited places, so I cantrack how much of my time is spend on which customer.

Use casesBased on our core users found from the user-stories we havecreated a set of use cases. The use cases provide crucial infor-mation about the context in which the prototype applicationwill be used.

Use case: The student Morten, 23, is really hungover fromlast night’s event and is trying to figure out what happenedand who he talked to during his escapades. An applicationwill quickly give him an overview of where the party startedand follow its progress place for place in a chronologicallyorder. On each event he can view who he interacted with boththrough texts and calls. Morten is now able to bind together astory of what went on last night.

Use case: The Elderly Jane, 62, has been diagnosed with aMCI[7] (Mild cognitive disorder), which causes her to forgetnew events. This is often apparent when going out shoppingon her own and talking to her children on the phone. Butafter a few hours she forgets who she has spoken to. Thechildren sometimes get irritated when she has phoned themfour times in one day. An application will be able to keeptrack of who she has phoned, how long the conversation wasand also where she was when the conversation took place.

Use case: The life loggist Jesper, 41, is logging every eventof his life and write down places he has visited in his diary.This is a very time consuming task to do, and he tends toforget to do it every day. The days’ events is piling up andin the end of the week he can not really remember what wenton at Monday. An application automatically tracks places hevisit and now he can with go back in time with a few clicks.The only problem is now, that all his old data is not as coollooking as the new collected data.

Use case: The consultant Lars, 34, usually spends a lot oftime on-site at different customers. Keeping track of wherehe has been and for how long, is manually written in an elab-orate spreadsheet to keep track of billing hours. Just as wellwith long client calls which are also billable. Keeping trackof all of this takes its toll on Lars and mistakes are boundto occur. An application which could log this data for Larswould greatly improve the billing accuracy to his clients, andsave him from stressing about remembering to log the datahimself.

Summary of our user stories and use casesWe became aware of our core users and in which situationsthey would use the application. We have extracted the follow-ing focus points. At this point only the students from DTUwill be able to use the prototype application.

• Where the user has been divided in days.

• Who the user has interacted with and where he has inter-acted with them.

• Simple and intuitive display which matches many usertypes from students, business people, elderly to enthusi-asts.

3

The core users have very different backgrounds, incentivesand needs. However, they still have a common goal; Quanti-fied Self[5]. Tracking the users movements and calls, display-ing the data in a simple way will make it easy for everyone tounderstand and interpret them.

PrototypingThroughout the development of the prototype we have madeextensive use of user-centered design principles so the finalapplication will satisfy and give the users a more user-friendlyuser experience - for a final product this will be the make orbreak on the market.

MockupFollowing the principles of user-centered design we wentthrough several iterations of mockups before finding the finaldesign for our prototype. An iteration consisted of a conceptwhich lead to a mockup, which was then tested and based onthe evaluation and feedback we revised the mockup. Eachmockup was evaluated on the students from the course andfrom their feedback we made improvement to the mockups.

The first mockup featured the use of the CardsUI[6] as thiswould give us opportunity to combine different types of ac-tivities. Each card would represent an activity i.e. a text,a phone call, a place visited and so on. The mockup is il-lustrated in Figure 6 and Figure 7. Our feedback from thismockup was that the interface would easily be cluttered whenadding more than a few activities and the list would be verylong.

Figure 6. Mockup 01 - CardsUI

The second mockup featured an offspring of the first one andbased on the feedback we got from the first one we decidedto change the layout. To solve the cluttering and long time-line, as noted in the first mockup, we introduced the cards ingroups and each card activity were assigned an unique color.The mockup is illustrated in Figure 8. The feedback that wereceived was, that it was less simple to look at, and the userswanted more content on the screen instead of having to scrolletc.

In the third mockup we skipped the CardsUI included in thefirst two mockups. The new mockup introduced a list of theactivities including their unique colors and this time also rel-evant icons. Skipping the CardsUI we made room for morecontent shown on the screen. The mockup is illustrated in

Figure 7. Mockup 01 - CardsUI - ActivityCard

Figure 8. Mockup 02 - ListView Colorcoded

Figure 9. The feedback we got from the focus group was thatthe mockup lacked a chronological timeline of the activities,but the listview was a great idea.

Figure 9. Mockup 03 - Activities in a listview

In the fourth mockup we tested out a radical change in thelayout, showing activity icons with content and a timeline be-tween each activity. This mockup was intended to only testthe timeline feature. The mockup is illustrated in Figure 10.The focus group loved the the timeline between each activitywhich gave a good separation between activities. One thingthey missed, however, was the listview from the last mockup.

Based on the feedback we decided to settle for a final designfor the prototype based on the third and fourth mockup. These

4

Figure 10. Mockup 04 - Timeline of activities

two mockups were the ones that the users were the most fondof.

JSON-parserTo make the design, the data has to be collected, digested andpresented in a standardized way. The data is collected froman url containing the address, access token and parameterswhich define what data should be collected. The process ofcollecting the data from the SensibleDTU-database is visual-ized in Figure 11.

Figure 11. Gathering the data

To get the relevant data, this process has to be done threetimes. One for gathering the locations, one for the call-logand one for SMS. With the help of the API, the data is col-lected back in time when getting locations and only on spe-cific date when retrieving call-log and SMS. Since SensibleDTU collects a lot of GPS locations every day the iteration ofretrieving the next url is imperative to collect all of the dataper day. For the SMS and call-log this is not a problem, it isour estimate that average people do not log more than a thou-sand texts or calls a day. The received format is json, whichis then easily iterated through to get either next link or theresults which are well structured. Every location data pointcontains an ID, TIMESTAMP and GPS-location that we areinterested in. When these are downloaded the process of fil-tering out duplicate locations is comprised of the followingsteps:

• Removing zeros from the GPS locations so accuracy isdown to the nearest meter. The GPS coordinates have moreprecise points but with some inaccuracy, so they are not re-ally that accurate, and unimportant this application.

• Removing data points within 10 meters. There will notbe many points of interests within 10 meters than can becomprehended.

• Removing data points where the user spends less than 15minutes which indicate that the user is on the move

When these steps has been taken, only a handful of locationsare presented per day. The SMS and call-log can then be gath-ered for the specific day and then divided out to the variouslocations.

RESULTSBased on the feedback from both the focus groups on ourmockups and the json-parser, we have developed two smallerprototypes each with its own purpose. The following sec-tion include each of these prototypes, a comparison againstmore comprehensible SensibleJournal and a user-testing onthe prototype application, SJournal.

PrototypeBased upon the feedback from the focus group on the earlymockups and the needs from our core users, we came up withthe final design for our prototype. The following are the fea-tures that we have included in our application.

• a complete summary of the day

• a chronological timeline

– including places visited

– including interactions with other people

• a simple and yet intuitive interface with unique icons andcolors

• a timeline indicator

In the top of each day we have added a complete summary ofthe day. The summary includes the total number of placesvisited and the number of interactions both on texts and phonecalls. Below the summary is the timeline for the entire day.The timeline includes each visited place throughout the dayin a chronological order. Each place is given a unique iconwhich indicates the type of place. We have included an actualtimeline indicating the time arriving at and leaving a givenplace. We have included the interactions with other peoplefor each place visited, so the user can look up interactions atspecific places. The list of every interaction could easily bea very long list and to solve this problem we have combinedtexts and calls based on the persons who the user has inter-acted with at that specific place. The timeline and summaryin the prototype is illustrated in Figure 12 and the interactionsbetween people for a given day is illustrated in Figure 14.

JSON-ParserThe data from the Sensible DTU server was collected andinterpreted with the implemented json parser. As Figure 13shows, the filtering narrows down to a few 100 locations andlater combining so only a few actual locations remains wherea user has been for more than 15 minutes. When the call-log and sms were collected, they were manually connectedwith views from SensibleDTU[10] to see that data could becombined with a location and the timespand at that place ofinterest.

5

Figure 12. Summary and timeline

Figure 13. Filtering tha data

Comparing SJournal with SensibleJournal 2013 and oth-ersThe Initial approach was to improve some of the aspectsdepicted in SensibleJournals. The areas that have been fo-cused on are locations throughout the whole day and howto make it more digestible for the user. By also adding calland text logs to locations, a more memorable experience willbe gained since locations will be connected with interactionswith friends or family. An example of this is shown in Figure14.

With the proof of concept accomplished, data can be collectedand digested. Static data was used from this to make sure wereceived consistent data to build the design. In this case theconnection with foursquare[14] was not implemented, thoughtests showed that the GPS coordinates gathered from SensibleDTU, coincided with actual locations on foursquare. Mainly

Figure 14. Interaction

focusing on locations and not movements, means occurrenceslike on a vehicle and walking is not in the scope as it is illus-trated in Figure 15 from SensibleJournal.

Figure 15. Interaction

Some of the design aspects were inspired by the various appsmentioned earlier. By making a timeline instead of a cardper day the user will get a more specific insight to activitiesduring the day.

6

User TestingIn order to test the final prototype we tested the prototype ondifferent users. There exists a wide range of different waysto evaluate mobile applications through user testing. The fol-lowing is a list of different testing methods which was con-sidered used:

• Recording observations

• Collecting usage data

• Benchmarking

• Participative evaluation

In this case, think-out-loud and heuristic evaluation has beenthe backbone of testing the design of SJournal. Participativeevaluation has been used in the process of eliminating lessfavorable designs.

One of the test participants for the thinking out loud evalua-tion, Marianne, 53, is used to computers and tablets, so shehas a healthy attitude towards using applications and designneeds. A small focus group of MCI-patients[7] at Glostruphospital has also viewed and commented on SJournal.

Marianne, was asked to open up the application and thenasked to give her first opinion about what the different ele-ments in the interface was. Immediately she spotted the time-line and told us that this was the path throughout the day. Aswe nodded she kept indicating the different elements. Hernext assignment was to find out how many texts and calls shereceived at a specific place. A little thinking out loud later shefound the specific place and clicked on it. Hereafter she gaveus the result for the question. The test with Marianne was asuccess and proved to us that the interface was designed in auser-friendly way.

The MCI patients were introduced to SJournal as a sideproject. Since they were also asking for a simple way to seewhere they had been and who they had interacted with, intro-ducing them to SJournal made sense. They particularly likedthe simplicity, since smartphones normally confused themwith all the different things happening everywhere. Some-thing ”simple and not too flashy” was their comment, andthey liked the connection between places and interactions.This also helped with understanding real world applicationsfor life logging.

Throughout this semester other students following the course,has given feedback in the form of peer reviews. This hassometimes helped steer away from bad design ideas and insome cases help guide in the right way.

Controlling the dataThroughout the project a task to write a diary on a daily basiswas given. The purpose of the diary served as a control-pointwhen handling the gathered data from the SensibleDTU API.One of the control diaries is included in GoogleDocs[12] anda snippet of it is illustrated in Figure 16. Besides using themanually diary, a control has also kept with Moves[1]. Thishas given a great view of how the data looks on a timeline,and this may have had an impact on how the data should bepresented in SJournal.

Figure 16. Snippet of Diary

Concerning control of the data the control-data from the di-aries has been used to pinpoint the most relevant datapoints inthe json-parser prototype, but as the main focus was to createa more user-friendly interface, the development of the data-prototype was discontinued.

DISCUSSIONDuring the process of the this project we have designed,tested and implemented two vertical prototypes - one for gath-ering the data and one for presenting it.

We started with a simple brainstorming and based on ourideas we came up with different mockups. We went throughseveral iterations and for each iteration we had feedback fromthe other course participants. Based on the feedback fromboth the course participants and our core users found fromthe user stories, we combined their needs and requirementswith the summary from our market analysis of existing ap-plications. With this in mind we were able to finalize ourinterface for our prototype while following the user-centereddesign principles.

The final prototype was tested and shown to a wide varietyof users with different backgrounds. We based our user test-ing entirely on a think-out-loud-interview and heuristic eval-uation. From the interviews we received great feedback andthe think-out-loud interview really showed that our interfaceis very user-friendly and intuitive, as Marianne did not haveany real problems solving the tasks given. The heuristic eval-uation from the MCI-patients showed that even elderly peo-ple could easily use the application, and even mentioned thatthere was a need for these kinds of applications for them.Even though the elderly is part of our core users, SJournalis not intended for their purpose in this project. On the otherhand we will not entirely decline the opportunity to developa similar application, which would fit their needs even more,as we can see a great market for these kinds of apps.

During the process we stopped working on the json-parser,since we had proved that the data could be collected, filteredand displayed. However, the process of downloading the dataevery time the application was run took a long time, since a lotof data points had to be collected. The data could have beenstored locally in files or on the application via a database.Since this was not the main focus of the prototype, we reliedon static data from our journals to illustrate the look and feelof real world data instead.

The time barrier stopped us from combining the two devel-oped apps. It would have been a great feature to be ableto show real world data in the application, so any studentenrolled in the SensibleDTU could have viewed their owncollected data. For translating the places (latitude/longitude)from the SensibleDTU API we could have made use of thepublic search API from foursquare[14], so real and user-friendly place-names were shown instead.

7

When creating the user prototype, we have considered the fivestages of personal information collection [13].

The preparation stage is a great part of the SensibleDTU-project as the great motivation behind this was to gain insightto how social networking look in real-time.

The collection stage was done by the DataCollector providedby the SensibleDTU, which collected all the relevant datafrom the sensors in the smartphone and then storing it in thedatabases on a specific frequency.

The integration stage comprised of downloading data fromthe Sensible DTU server and filtering out irrelevant datapoints and only showing actual visited locations in a simpleconstruction.

The reflection stage then presents the data in a simple com-prehensible way.

The action stage has been overlooked since the projects focuswas to present the data and not suggest any future activitybased on the presented data.

CONCLUSIONDuring the project we have developed two vertical ap-plications which individually have their own benefits andstrengths. Combining the prototypes we would end up with aproduct that could lead to a final Android-application whichcould be released on the market after some further develop-ment.

By using methods like user-stories and use-cases we gainedknowledge and a greater overview about who our core userswas, what features they needed in the application and in whatdifferent types of context the applications would be used in.

We made extensive use of the user-centered design principlesin order to get to a final prototype, which we then could teston different test participants. Our tests showed that the hardwork on going through several iterations of the design-phasepaid off. The test participants liked the idea of being able toget a timeline of who, when and where they have interactedwith people in their lives.

The tests also confirmed the potential of these kinds of appli-cations being more and more interesting for the general popu-lation. Quantified self is here, and in the next couple of yearsthe devices allowing lifelogging will be more widespread.

Our vision for this project was to design, test and developan application that could improve the interface of the currentSensibleJournal-application.

Future WorkBy combining the two prototypes users will be able to actu-ally see where they have been. By implementing a databaseit would be easy to save, structure and maintain data pointsthroughout the day.

The data should be crowd sourced by letting the users editnames of places they have visited. By calling their homefor ”home” and DTU library for what it is, instead of ad-dresses on a road, which could be rather impersonal for users

and tedious for them to remember. This, combined withfoursquares location names, will help the user feel a closerconnection and easily remember previous interactions at lo-cations.

A dropdown button on the top to change between the variousdays should also be present so the user can look further backin time to remember past activities. An overview of how longthe user has spent at home, work or DTU could also give anice look back at the last week, and perhaps motivate them toprioritize their time differently.

Individual contributionsDuring the process of the project we have divided the work-load equally and the following is the sections that each of ustook responsibility of.

Ulrik• The graphical prototype

• Abstract

• Introduction

• Related work

• Results

• Conclusion

• Team spirit

Kristian• The json-parser prototype

• Abstract

• Methods

• Discussion

• Conclusion

• Team spirit

REFERENCES1. Moves, Activity Tracker Without Gadgets,

http://www.moves-app.com, (02 Dec 2013 at 12:54)

2. Friday, A storyline of your life,http://www.fridayed.com, (02 Dec 2013 at 13:14)

3. Trails, Trails mark your life experiences into a map,http://trails.fridayed.com, (02 Dec 2013 at 13:26)

4. SensibleJournal, TThe look of a social network in realtime, https://www.sensible.dtu.dk/?lang=en, (02 Dec2013 at 15:46)

5. Quantified Self , A collaboration of users and toolmakers who share an interest in self knowledge throughself-tracking, http://bit.ly/1bIgu7X, (04 Dec 2013 at15:36)

6. Google CardUI, Google Card UI (extracted from GoogleNow, https://github.com/afollestad/Cards-UI, (04 Dec2013 at 17:36)

8

7. MCI, Mild cognitive impairement,http://www.alz.org/dementia/mild-cognitive-impairment-mci.asp, (05 Dec 2013 at12:36)

8. Lifelogging - whats next , Capturing and sharing alldaily minutiae will soon get automated. Very soon.,http://bit.ly/1eYGQ4j, (05 Dec 2013 at 12:36)

9. The Lifelogging Trend, What Can We Expect From It?,http://doktorspinn.com/2012/08/18/the-lifelogging-trend-what-can-we-expect/, (05 Dec 2013 at12:58)

10. SensibleDTU, a large interdisciplinary research project,involving researchers from several departments at DTU,http://bit.ly/sensibledtu, (05 Dec 2013 at 13:15)

11. Prototype Slides p. 20, Slides from 02827 MobileApplication Prototyping, http://bit.ly/1g9wfGP, (05 Dec2013 at 13:51)

12. Sensible Diary, Day to day diary of our whereabouts,http://bit.ly/IR89mC, (05 Dec 2013 at 14:57)

13. Prototype Slides p. 20, Slides from 02827 MobileApplication Prototyping, http://bit.ly/1cnW3eW, (05Dec 2013 at 16:52)

14. foursquare API, The foursquare API gives you access toall of the data used by the foursquare,https://developer.foursquare.com, (10 Dec 2013 at10:52)

15. Jawbone Up, UP is a wristband and app that tracks howyou sleep, move and eatthen helps you use thatinformation to feel your best.,https://jawbone.com/up/international, (10 Dec 2013 at10:13)

16. Fitbit, Track your steps, distance, and calories withFitbit, http://www.fitbit.com/uk, (10 Dec 2013 at 10:13)

9