Progress Highlights
NOTE: This DevLog is a combination of my progress for both Week 7 and Week 8
Project Research
Obtained approval from supervisor for final version of Table of Comparison
Split master Table of Comparison (in Excel sheet) into 2 tables in Word document (Related Works on Detecting Mental Health on Twitter, Related Works on Detecting Mental Health on Bilingual Twitter Posts)
Updated HTA diagram (added and updated functions under existing modules, split Expert and Tweets Set Management into separate modules)
Updated class diagram (removed certain classes due to redundancy and updated existing classes related to Labelling System with concrete function implementations)
Updated list of requirements (functional, quality, constraints) and use case descriptions with Faidz based on comments from supervisor
Updated use case diagram based on comments from supervisor (removed redundant use cases, added use case ID inside naming)
Created a simplified version of deployment diagram using Visual Paradigm (following n-tier architecture pattern)
Created entity relationship diagram (ERD) for the web-based systems of Fitweet
Drafted UI/UX design or web app design mock-up for the Labelling System using Adobe XD (had to intentionally make it slightly hideous in order to not spoil the finalised design of the deployed web-app)
- Prepared presentation slide and recorded a 15-minute presentation for FYP1 monitoring 
Project Development
Monday (19/4/2021):
- Added paginations for Expert Management table
- Configured FBV to include Django paginators code for Expert Management
- Added Django paginators template tags in Expert Management template
- Updated settings.py to include SESSION_COOKIE_SECURE and CSRF_COOKIE_SECURE variables
- Updated settings.py to include dynamic ALLOWED_HOSTS, DATABASES, STATICFILES_DIRS and STATIC_ROOT values based on DEBUG
- Removed SESSION_COOKIE_SECURE and CSRF_COOKIE_SECURE values due to absent of SSL certificate
- Fixed bug for incorrect Tweet ID inside certain Tweets data sets that causes 500 server error from labellers (replaced exponentiated Tweet ID into full integer form)
*Thursday (22/4/2021):
- Changed redirect logout to logout HTML for administrator and updated Instructions for Tweets Labelling
- Added multiple tweets set file upload functionality
- Added Django paginator functionality for Tweets Set Management
- Added Download Expert Details and removed auto-download for bulk experts creation in admin.js
- Fixed bug for swapped columns (Assigned Set, Full Name) in Bulk Experts Generate
Monday (26/4/2021):
- Updated Instructions for Tweets Labelling page per supervisor's request
- Added order_by function to sort generated expert report based on set name
- Added order_by function to sort set and expert management based on set name
- Fixed bug on set name sort for Experts Management
- Fixed bug on set name sort for Experts Management
- Added metatags for Log Out
- Renamed all Log Out instances to Sign Out
- Renamed Log In to Sign In
What I'm stuck at?
Since I'm done with the FYP1 monitoring presentation recording and the experts labelling phase is halfway through (14 days period was given to them), currently I'm not stuck at anything in particular.
How will next week be?
- Compilation of all labelled Tweets set by Experts
- Meeting with supervisor to discuss next course of action (ML model development)
Lessons Learnt
Never assume that your work is correct on first try. Always try to get secondary external opinions to validate the correctness and quality of your work. For instance, I always try cross-check the work I've done with my partner and supervisor to ensure that I'm on the right track. It can be annoying and tedious at first especially when you have to repair them frequently, but it will be worth it in the end.
No system is bug-free. After the deployment of the labelling system and its official release for the Experts to use, we encountered numerous bugs. This is to be expected since its release was rushed and we didn't have time to do proper testing. In fact, seniors have advised me to be careful of deployment bugs and never be too confident that everything will work in first try. Some bugs were minor and easily fixable while some requires in-depth analysis. Thankfully, we managed to fix most of them so far and until the labelling phase is complete, I have to be on high alert for possible errors/bugs reported by the Experts.
Strive to do more than the bare minimum. What differentiates between you and your competition is on how far are you willing to exceed expectations. For the FYP1 monitoring presentation, initially my supervisor advised that we only present the summary of the literature preview in only 1 slide. However, knowing that this might not be enough especially considering that our FYP is research-based/research-heavy, I asked if it's possible for us to include more content. She gave a green-light provided that we can present within 15-minutes (which I fortunately managed to do so). Furthermore, both Faidz and I also brainstormed the possible ways in which we can make our presentation stand out, such as by including the collaboration outcomes among our stakeholders.