PSPPasay Scholarships · Pasay City · 2025 · 02/03

Giving Pasay scholars their time back

Pasay City's ₱20 million scholarship program ran on paper forms, long lines and four Excel workbooks. We built a web platform to move it online: applications, pre-filled renewals, OCR-assisted document checks and role-based approvals. It passed every functional test and scored 4.74 out of 5 with 30 testers, then stalled at handover.

built for
1,700scholars per semester
user testing
4.74out of 5, 30 testers
functional tests
7/7modules passed
architecture · 7 stepsfollow the request in §04
  • Scale: 1,500 to 1,800 applications per cycle and about 1,700 active scholars, with renewing scholars re-entering the same data every semester.
  • My role: I pitched the idea, then led delivery and quality: the test cases, the fix list, internal deadlines, user acceptance testing, and most of the final paper. Our team lead led development, and our Product Owner handled communications with the bureau and helped with documentation.

Context and constraints

The program funds Pasay high school and college students who study outside the city, under strict rules: proof of residency, a general weighted average of 80% or higher, and no failing, incomplete or dropped subjects. New applicants submit eight documents; renewing scholars submit three each semester.

  • Sensitive data under law: IDs, grades and indigency certificates fall under the Data Privacy Act of 2012 (RA 10173).
  • Short, fixed windows: applications open twice a year for about a month.
  • Four kinds of users: administrators, bureau staff, applicants and guests, each with different access.
  • Uneven digital access: some applicants have limited internet or digital skills, so the interface had to stay simple and work on phones.
  • Clear scope edge: the system ends at the approve-or-deny decision; payroll and fund release stay outside it.
  • Real team capacity: three students, all scholars ourselves (DOST and a city scholarship program), each also carrying leadership roles in other organizations.

Listening first

The team interviewed 2 bureau staff and 6 applicants before designing anything, and their answers decided the feature list.

What we heardWhat we built
Status updates by SMS were getting flagged as spam, so applicants missed themIn-app status tracker and notifications
Applicants re-entered the same personal details every semesterRenewals pre-filled from last semester's approved record
Applications were rejected for missing or wrong files, discovered only at the counterEligibility flags raised the moment a record is saved, with error-specific re-upload prompts
Staff insisted certified true copies still need a human checkOCR assists reviewers instead of auto-approving; every award has a named approver
Applicants were unsure of deadlines and requirementsPublic Explore page with program details, open to guests

Before and after

Data is captured once, at the source, instead of being re-typed by staff at every step.

  1. Before, on paper: Posts on the bureau's Facebook page

    Announce

    After, on the platform: Explore page with program details and requirements

  2. Before, on paper: Paper form plus eight documents, in person; incomplete sets turned away

    Apply

    After, on the platform: Online application with uploads from any device

  3. Before, on paper: Staff check completeness by hand

    Check

    After, on the platform: Eligibility flags on save, plus OCR to help verify uploads

  4. Before, on paper: Re-typed into four Excel workbooks

    Encode

    After, on the platform: Entered once into a single database

  5. Before, on paper: Scholars re-enter personal details every semester

    Renew

    After, on the platform: Renewal pre-filled from the existing record

  6. Before, on paper: No visibility; applicants ask in person

    Track

    After, on the platform: Real-time status tracking with notifications

  7. Before, on paper: Paper folders routed for signature

    Approve

    After, on the platform: Role-based approval workflow; two-factor sign-in for approvers

  8. Before, on paper: Manual tallies in Excel

    Report

    After, on the platform: Excel, Word and PDF exports generated by the system

The system

Our team lead built the platform; this is how it works, from upload to award. Rules live in the database, so the same checks apply whichever screen a record comes from. Follow the request:

Every award passes an eligibility check and two human reviews yes no fix and resubmit Apply online OCR reads uploads Complete andeligible? Staff review Returned with reason Admin approvaltwo-factor sign-in Decision and reports
Every award passes an eligibility check and two human reviews. 7 steps follow.
  • Eligibility on save: a database trigger flags missing documents, voter status and academic standing whenever a record is created or updated.
  • Renewals: database functions create next semester's record from the scholar's last approved one, then re-check eligibility.
  • Access by role: row-level security lets applicants touch only their own records, lets staff read and update but never delete, and leaves global settings to admins.

My functional test cases covered the flows built on top of these rules: registration, login, profiles, application submission, status tracking, notifications and session handling.

Key decisions

OCR assists; people approve

choice
OCR still reads uploads, especially application forms, but the platform and the reviewers decide together: software checks the rules and flags what's missing, and people judge the documents and approve each award.
instead of
Letting OCR approve matching documents automatically.
why
Bureau staff pushed back: certified true copies still need a human check, because that keeps an award legally sound. It's slower than full automation, but every award can be traced to the person who approved it. This is the decision I'd point to first.

Rules live in the database

choice
Eligibility checks run as a database trigger, renewals as database functions, and access as row-level security.
instead of
Checks written into each screen.
why
The same checks apply whichever screen a record comes from, and staff can read and update records but never delete them.

Renewals start from last semester

choice
Each renewal is pre-filled from the scholar's last approved record, then re-checked.
instead of
Scholars re-entering their personal details every semester, as they did on paper.
why
It came straight from the applicant interviews, and it gives the people this was built for their time back.

Turning it around

Testing came too late in our plan, and it showed: a long list of issues surfaced in the final stretch, because building and testing had been packed into the last weeks with no buffer. As a team, we reset how we worked, and I took ownership of quality and delivery.

  1. One fix list. I turned every issue into a tracked item and closed the list out one by one, so nothing depended on memory.
  2. Written test cases. I wrote functional test cases for all seven modules, from registration and login to status tracking, notifications and session handling.
  3. Internal deadlines. I set and tracked checkpoints so features reached testing well before review dates, not the night before.
  4. Real users, real devices. I ran user acceptance testing through a structured questionnaire, validated by our adviser and the Pasay bureau. Thirty testers from nine cities used the live build on their own phones and laptops for 10 to 20 minutes each.
  5. One document. I wrote most of the final paper, kept it in step with what the system actually did, and presented it with the team at the defense.

The result: every functional test passed, and testers rated the platform 4.74 out of 5.

Outcome

The study concluded the platform was ready for implementation.

  • Functional testing: all seven modules passed, from registration to session handling.
  • User acceptance testing: 30 testers from nine cities: students, teachers and administrative staff, including staff from Pasay City. All seven categories landed in the top band, from 4.66 for interface and performance to 4.81 for overall evaluation.
  • Where it stands: the study recommended the bureau adopt the platform as its official system. It has not been deployed. We reached out to the bureau after the defense, but our messages went unanswered, and the code now sits in a repository, working and unused. A prototype build is still online, but I don't link it here, because it carries the city's branding and staff contact numbers. There was no handover plan and no one inside the bureau to carry it forward, and after a year-long marathon our team was too tired to push harder.

The platform was built for a program of about 1,700 scholars per semester.

What I'd do differently

Three rules I'd follow next time:

  1. Plan for real capacity, with clear roles and expectations from week one.
  2. Make testing a gate, not a phase, and share code ownership so no one person becomes the bottleneck.
  3. Plan the handover before building, with a named owner on the client's side.

The platform itself would grow the same way the study suggests: check residency and grades against city and school records instead of uploads, design for phones first (interface and performance was the lowest-scored category), and add a Filipino language option.

The full story, and advice for students starting their capstone, is in the note You Don't Have to Write the Code to Be the Reason It Works.

Credits

Built by a team of three: our team lead (Scrum Master, lead developer), our Product Owner (communications and documentation) and me (delivery and QA lead), with our faculty adviser, De La Salle University College of Computer Studies. Drafted from our capstone paper, CAP-2412-IT (Nov 2025).

end of psp.pasay-scholarships.md. next: ~/work

~/angelo · Metro Manila · © 2026

privacy

This site sets no cookies and runs no ads or trackers. Cloudflare hosts it and counts page views with Web Analytics, which uses no cookies; a counter of my own records only how many times the CV is downloaded, nothing about who. Your theme choice, open tabs and the notes you pinged stay in your own browser. If you send a signal under a note, Cloudflare Turnstile checks that you're not a bot; the site keeps only the count. I collect no names, emails or IP addresses. For questions or requests under the Data Privacy Act of 2012 (RA 10173), write to angelo@nuque.dev.

~/angelo/work/psp.pasay-scholarships.md

keyboard

move

j / k
scroll down / up
d / u
half a page down / up
gg / G
top / bottom
] / [
next / previous section

go to

gh
README.md (home)
gw
work/
gn
writing/notes/
ga
about.md
gt / gT
next / previous open tab

panels

: / /
command palette and search
ctrl+k
command palette (always on)
esc
close, or leave :zen
?
this list
:tour
a 30-second walkthrough