IS480 Team wiki: 2015T1 Vulcan Final
|HOME||ABOUT US||PROJECT OVERVIEW||PROJECT MANAGEMENT||DOCUMENTATION|
- 1 Project Progress Summary
- 2 Project Management
- 3 Schedule Metric Formula: (Estimated Days / Actual Days) x 100%
- 4 Quality of product
- 5 Reflection
Project Progress Summary
Highlights of Project:
- LiveLabs relocation of servers on Oct 5th 2015, two days before Midterms Presentation
- LiveLabs servers overload (data increase 1GB/minute), seemed to have been caused by us
- API level of phone borrowed from school too low for our development
Please refer to Planned vs Actual Tasks Metrics for the detailed breakdown of our individual tasks.
Project Schedule (Plan Vs Actual):
Schedule Metric Formula: (Estimated Days / Actual Days) x 100%
|Iteration||Planned Duration (Days)||Actual Duration (Days)||Schedule Metric Score||Action||Status|
|2||18||32||56.25%||Team is behind schedule. This is due to the complexity of the tasks planned (Android App and Smart Watch).
Follow up action: Rescheduled the future iterations, deducted days from buffer days.
|4||18||24||75%||Team is behind schedule. This is due to Livelab's server permission issues.
Follow up action: Rescheduled the future iterations, deducted days from buffer days.
These are the top risks we have identified and has happened before Mid Term. We have followed the mitigation steps listed above and successfully managed the risks.
Beeper Survey Creation:
Beeper surveys are created each day by an alarm manager using the set method. The setWindow method was initially used as it would schedule the alarm within a given window of time. However, this did not ensure that participants received beeper surveys at a random times because the beeper surveys would go off close to the start time. In order to fix this problem, the randomisation of beeper timings was done beforehand through java. This was done because there was no method to implement what the sponsor wanted which was 3 random beeper surveys each day with the timings being all different.
There will be 3 blocks of equal duration created based on the wake up time and sleep time which the participant has inputed. From there, 3 random numbers will be generated with the max number being the length of a block. These 3 numbers will be added to the start time, block 1 time and block 2 time to create the 3 random times for the beeper surveys.
Android Studio's alarm manager does not allow repeating alarms to have a different time everyday and thus this could not be used. The solution from our team was to create a repeating alarm that will go off at midnight if the participant's sleep time is before midnight and at the sleep time if the participant's sleep time is after midnight. The repeating alarm will call the BeeperCreator class which creates the 3 random beeper survey timings for the day.
Dynamic generation of survey elements:
Normally to implement dynamic lists of content, we would use the Android ListView to display the elements on the activity page. ListView is useful for creating a scrolling list of elements that can be interactable and is efficient in terms of implementation. However, due to how ListView recycles elements that move out of view when scrolling, it is difficult to retain user entered information that is linked to an element, as the information would be inherited by the "new" element that comes into view.
To avoid this, we implement ScrollView instead. ScrollView is useful for showing a scrolling page of static elements but its not usually used to generate a dynamic list. The above shown implementation indicates how we append a layout "fragment" to the current ScrollView to represent a single survey question element. We populate the question fields for each question element and attach it to the view, allowing a scrolling view that will display questions according to what is required for this particular survey. All question elements are active in memory and are set up to collect user input as indicated in the implementation below.
As can be seen from the given Logical Diagram, our database implementation includes tables which are dynamically generated whenever a new study is added by a researcher. The tables reflect the given name of the new study and collectively contain the survey configuration and result data for that study. This isolates the data collected for each study so that they do not interfere with each other or potentially mix. Furthermore, results are separated into different tables such as "rfk_beeper_result_*program name*" and "rfk_session_result_pause_*program name*", where "program name" is the name of the study. This facilitates the separate retrieval of data of different types by the researcher and allows future development to cut or add more data types without having to modify the tables.
Quality of product
Details of users, specifically login details which could be personally identifying, are separated from demographic information about the user and their result data. Therefore, in the event that a study participant requests that they be removed from the program along with all identifying details, they can be removed from the database while retaining their studies data as anonymous participants, protecting their privacy.
Furthermore, in order to ensure that user data is not proliferated, only the creator of a study is able to access a study and modify its details, and more importantly retrieve result and demographic data about participants in that study. Other researchers are unable to access other studies, which if made possible may be a breach of privacy as participants may have only provided permission to results to the owning researcher. For administrative purposes, any user with administrator rights will also be able to access all studies and their data, as representatives of the Refokus system.
In order to provide scalability and flexibility to researcher created studies, creation of a study can include the creation of an unlimited number of text or slider based questions for post-session and periodic beeper surveys. This gives researchers the ability to customize their surveys to a large extent and it potentially accommodates any possible data points the researcher may wish to collect through survey data.
Furthermore, updating of survey questions or session-specific podcasts is possible while the study is active where there are existing users with partial progress. Completed sessions that receive updates will not be repeated for participants, but any sessions they have not yet completed will be updated to the latest attributes set by the researcher. Any data collected will reflect the survey results according to the version completed by the participant, and so no data is lost either from the old or new version of the study.
In order to improve availability of service, users will be able to download the podcast for their next session ahead of time, even if they are not yet able to start the session due to the imposed daily limit. In the process the mobile app will also update its survey questions to be used for any subsequent post-session survey or beeper surveys. Sessions can be carried out without internet availability, and the result data stored after the session until internet connectivity is restored.
|Project Management||Minutes (Sponsor,Supervisor, Team)||Minutes|
|Metrics||Schedule, Bug, Change Management Metrics|
|Requirements Gathering||Design Documents||Scenario,Storyboard,Navigation Diagram,Prototype|
|Market Research||Market Research|
|Analysis||Use case||Participant & Researcher Use Cases|
|Business Process Diagram||Business Process Diagram|
|Logical Diagram||Logical Diagram|
|Testing||User Testing||User Test|
|Test Plans||Test Cases|
We have deployed our mobile application on the Google PlayStore:
|Alpha Version||Instructions to download|
|Beta Version||Instructions to download|
|Production Version||ReFokus is on PlayStore now!|
We have deployed our web application on the Livelabs Servers:
|ReFokus on Test Server|
|ReFokus on Production Server|
We have also deployed our web application to the Livelabs Web Server (Hestia), ReFokus Web Application
Number of User Tests: 3
Our testers consist of users with research backgrounds, specifically research assistants currently pursuing their PhD in Psychology. With their experience, we were able to gain valuable feedback with regards to the creation of studies.
For more information about the user tests and the detailed results, please visit the link below:
For each iteration, we have functional test cases to test individual functions. Towards the end of the iteration, we will do regression testing and go through the entire flow of the project to ensure all parts are working.
For the detailed test cases, please visit the link below:
For our bug metrics score, we can see that iteration 5 had an exceptionally high score. This was due to the aftermath of User Test 2 and 3, which proved to be useful for us with the functional and UI bugs spotted. Even though the bug score was well above the threshold level of 10, we managed to solve all the bugs in the scheduled debugging time.
For the detailed bug reports, please visit the link below: