Data Sources — ThuSh LK

What's in this document

Tap any section above to jump straight to it

Tab 01 · Sources

Six Pillars of Verified Truth

The six primary source types — University Handbooks, Official Websites, UGC Publications, Research Papers, Direct Calls, and Email Correspondence.

Tab 02 · Method

Seven-stage Research Pipeline

How raw sources are turned into trusted content — collection, cross-referencing, direct confirmation, structuring, review, publication, and continuous monitoring.

Tab 03 · AI Risk

AI Displacement Risk Calculator

Sri Lanka's first degree-level AI automation risk score — composite formula, seven research frameworks, four risk bands, and the build pipeline.

Tab 04 · Tools

All Platform Tools & How They Were Built

Every tool on ThuSh LK — Z-Score, Degree Finder, Comparison, AI Risk, Cut-Off Tracker, Cart — with the methodology behind each.

Tab 05 · Vision

The Story & Integrity Commitments

The founder's four-year journey, our public integrity promises, and how to report an inaccuracy.

Pillar 01

University Official Handbooks

Verified

The latest edition of each university's official student handbook — the authoritative document published by the university itself. Every course structure, credit count, and graduation requirement is traced back here.

EditionLatest only

FormPrint + PDF

AuthorityFaculty Board

Pillar 02

Official University Websites

Verified

Public academic portals of every UGC-approved university — faculty pages, department announcements, programme pages, and official PDFs. Tracked continuously; every change is logged.

Coverage17+ unis

Check cycleWeekly

ArchiveVersion log

Pillar 03

UGC Handbook & Publications

Verified

The Admission Handbook, cut-off tables, district quota allocations, and all official publications of the University Grants Commission of Sri Lanka — the statutory authority on admissions.

PublisherUGC SL

RefreshPer A/L cycle

RecordsHistoric + now

Pillar 04

Research Papers & Studies

Verified

Peer-reviewed academic studies on Sri Lankan higher education, graduate employability, curriculum analysis, and workforce trends — used to inform career outlook, degree outcomes, and the AI automation-risk model.

ScopeLocal + global

UseContext analysis

CitationOn each page

Pillar 05

Direct University Calls

Direct

When published documentation is unclear, outdated, or contradicts itself, we call the faculty or department directly. Registrar's offices, academic coordinators, and programme heads — person-to-person — to confirm what is actually taught.

ContactsFaculty / Dept.

LoggedDate + role

TriggerAny conflict

Pillar 06

Official Email Correspondence

Direct

Written confirmation from authorised university channels — a paper trail for every disputed detail. Significant curriculum updates or non-public data points are backed by an email thread with a named university staff member.

FormUni domain only

RetainedInternal archive

UsageCross-check

Collection

Primary source collection

Latest official handbooks, faculty websites, and UGC publications are acquired and archived for the degree in question. A working dossier is assembled with every raw reference linked to its exact source.

Inputs · 3–6 docsOutput · Raw dossierOwner · Research lead

Cross-reference

Multi-source verification

Each data point — subject code, credit value, entry requirement, cut-off score — is checked against at least two independent sources. Aligned data is marked verified; divergent data is flagged for direct confirmation.

Rule · ≥ 2 sourcesConflicts · LoggedStatus · Verified / Flagged

Direct Confirmation

Call or email the faculty

Every flagged item triggers a phone call or formal email to the relevant faculty, department, or academic coordinator. We record who we spoke to, their role, and the date of confirmation — not just the answer.

Channel · Phone + EmailLogged · Contact + dateTrigger · All conflicts

Structure

Structured data modelling

Verified content is mapped into the ThuSh LK framework — a schema covering specialisations, subject combinations, entry pathways, and career paths. A three-level classification (Domain → Interest Area → Specialization) makes 500+ courses directly comparable.

Schema · Research v2Fields · 60+ per degreeConsistency · Enforced

Review

Editorial & domain review

The structured entry is reviewed for clarity, completeness, and student-friendly language. Technical terminology is explained. Ambiguous wording is rewritten without changing meaning — the original preserved.

Check · Language + accuracyArchived · OriginalAudience · A/L students

Publish

Publication to the platform

Only at this stage does content reach a student. Each published degree page is stamped with its last-verified date and retains a version history so changes are transparent and traceable.

Stamp · Verified dateHistory · Full logAccess · Free forever

Monitor

Continuous monitoring

Publication is not the end. Sources are re-checked on a rolling cycle, and any student-reported discrepancy is investigated within our stated response window. Every correction is a correction of record.

Cycle · RollingReports · Student-drivenOutcome · Public correction

Accuracy standards

The measurable discipline behind every page

Source Rule

≥ 2sources

Every substantive data point must be corroborated by at least two independent sources before publication.

Handbook

Latestonly

Only the most recent university handbook is used as primary reference. Superseded editions archived, never cited as current.

Escalation

100%

Every flagged conflict between sources is escalated to direct faculty contact. No conflict is ever silently resolved.

Correction

≤ 14days

Target from a reported inaccuracy to a published correction. Most resolved much sooner.

Data Freshness · Update Cycle

Live

Daily

Inbound Reports

Student corrections triaged the day they arrive.

Weekly

Website Checks

Scheduled sweeps to detect silent faculty changes.

Annually

UGC Cycle

Full refresh aligned with A/L cycle.

On-release

New Handbooks

Any new handbook triggers a re-audit.

Composite Formula

The AI Displacement Risk Score

Weighted · Task-based · Locally calibrated

R_degree = Σ_i w_i · F_i · T_i + α · C_SL − β · H_human

RComposite risk score (0–100) for the degree

wᵢWeight per framework (peer-review, recency, SL relevance)

FᵢFramework i's automation probability output

TᵢTask time-share within the role

C_SLSri Lankan industry & economy adjustment

HHuman-interaction / judgement protection factor

α, βCalibration constants per degree domain

ΣSum across seven frameworks (i = 1 … 7)

Seven research frameworks

The base layers of the composite score

Framework 01Foundation

Task-probability Model

Classic probability-of-computerisation scoring of occupational tasks — the foundational layer of any credible automation model.

Framework 02High

Task-based Substitution

Decomposes roles into atomic tasks and estimates what share of each is technically automatable with current and near-term AI.

Framework 03Med–High

Global Workforce Outlook

International projections on disappearing versus emerging roles across industries and the net displacement/creation balance.

Framework 04Medium

Skill-demand Shift

Rate at which skill demands within a profession change — captures role-value degradation even where the role survives.

Framework 05High

Cognitive-Routine Axis

Classifies tasks on routine/non-routine and cognitive/manual axes — strongest predictor of near-term displacement.

Framework 06Recent

LLM-Exposure Index

Share of tasks exposed to modern large-language models specifically — the newest and fastest-moving pressure.

Framework 07SL Context

Local Labour-Market Model

Sri Lankan employment data, public-sector absorption, and industry maturity factors — the context layer global models omit.

+ ShieldSubtractive

Human-interaction Factor

Protective factor for empathy, judgement, care, and trust-driven tasks — least displaced even at high AI capability.

◉

Four risk bands

How the 0–100 score translates to student guidance

0 – 25 · Low

Low Risk

High human-interaction, judgement-led, or craft-intensive roles. Automation pressure limited within a decade.

26 – 50 · Moderate

Moderate Risk

Mixed task profile. Portions automate; core professional judgement remains. Reskilling expected mid-career.

51 – 75 · High

High Risk

Routine-cognitive share is substantial. Significant task automation likely within 5–10 years; role redefinition required.

76 – 100 · Severe

Severe Risk

Most tasks highly substitutable by near-term AI. Strong reskilling pathways or specialisation pivots essential.

AI Risk Build Pipeline · ~200 Courses

Reproducible

Step 01

Role Mapping

Each degree mapped to 8–12 representative job roles.

Step 02

Task Decomposition

Each role split into atomic tasks with time-share.

Step 03

Framework Scoring

Tasks scored across all 7 frameworks; normalised.

Step 04

Composite + Export

Aggregation, SL calibration, Excel database.

🎯

Tool 01 · Admission Live

Z-Score Calculator

The A/L starting point every SL student needs

Computes a student's Z-score eligibility against every UGC-approved course by stream, district, and year — the first checkpoint before any degree shortlisting.

How it was built

Historical Z-score cut-offs extracted from UGC Admission Handbooks year by year, cross-verified against published university admission lists. Built as a live lookup so students can test hypothetical scores against real historic thresholds.

17+

Universities

All

Streams

Yearly

Refresh

🗺️

Tool 02 · Admission Live

UGC District Determination Tool

Resolves the exact admission district — gets this right when others don't

Determines the exact district a student is classified under for admissions — the trigger for how their district quota is applied. Often misunderstood; often gets students wrong answers from less careful sources.

How it was built

Built directly from UGC rulebook logic — school location hierarchy, registered residence, and every special-case handling. Logic verified against multiple UGC publications, edge cases confirmed via direct cross-checks.

UGC

Source

All

Edge Cases

Yearly

Review

🔍

Tool 03 · Discovery Live

Smart Degree Finder

Personalised match, not just a list

A ten-step intake maps a student's subjects, interests, strengths, and aspirations to the degrees that statistically best suit them — not an alphabetical list.

How it was built

Each degree is encoded as a vector across six dimensions — Specialisation Depth, Career Breadth, Practical Exposure, Academic Rigor, Flexibility, Entry Breadth. The student's 10-step answers form their own vector. Matches ranked via cosine similarity, filtered by Z-score eligibility.

Steps

Dimensions

500+

Degrees

⚖️

Tool 04 · Decision Live

Smart Degree Comparison

Seventeen-section side-by-side — no cherry-picking

Compare any two (or more) degrees section by section across seventeen distinct facets — from core subjects and specialisations to career paths, AI risk, and practical exposure.

How it was built

Every degree structured to the same 17-section schema during the research pipeline, so each comparison is like-for-like rather than cherry-picked highlights.

Sections

Any

Degree Pair

One

View

🤖

Tool 05 · Future-proofing Live

AI Displacement Risk Calculator

Sri Lanka's first — composite across 7 research frameworks

A composite 0–100 automation risk score per degree, built on seven established research frameworks and calibrated for local labour-market conditions.

How it was built

Full methodology in the AI Risk tab. In short: each degree mapped to representative job roles → each role decomposed into tasks → each task scored across seven frameworks → results aggregated via the composite formula → adjusted for Sri Lankan context. Delivered as a research-grade Excel database built through an automated pipeline across ~200 courses.

~200

Courses

Frameworks

Risk Bands

🛒

Tool 06 · Workflow Live

Degree Cart

The bridge between browsing and deciding

Save any degree across the platform to a personal cart, build a shortlist, and carry it into the comparison tool — the connector between discovery and decision.

How it was built

Simple but deliberate: the cart is the workflow connector between discovery, filtering, comparison, and saving. Designed so students keep control of their shortlist without losing progress between sessions.

Cross

Page Sync

Any

Degree

Free

Always

📈

Tool 07 · Reference Live

Cut-Off Marks Tracker

Year-over-year trend lines — not single numbers

Year-over-year cut-off marks for every stream and course, visualised as trend lines — the shape of a course's competitiveness over time, not a single number in isolation.

How it was built

Data extracted from historic UGC Admission Handbooks, reconciled against published admission lists, and normalised into a single structured dataset. Each trend is a small but trustworthy story a student can read in seconds.

Multi

Year History

All

Streams

UGC

Verified

★ From the Founder

A student's future should never rest on a rumour. Everything on ThuSh LK exists because I could not find it anywhere else — clear, honest, and in one place.

ThuSh Pramod

Founder · Researcher · Conceptualiser

● The Story

ThuSh LK was conceived, structured, and is personally maintained by ThuSh Pramod — a Sri Lankan government university graduate who spent his own years inside the system the platform now documents. Every category, every framework, every comparison tool on this site began as a question he was asked, could not quickly answer, and refused to leave unanswered.

Over four to five years of continuous research, that effort grew into Sri Lanka's most comprehensive independent higher-education information centre. The platform's underlying structure — the three-level degree classification, the subject-to-specialisation mapping, the six-dimension degree scoring system, the seven-framework AI risk composite, and the district quota logic — is entirely his own design.

The concept, the thinking patterns, and the mental models behind every tool on this site belong to one person. The goal has always been singular — a Sri Lankan student should be able to find the truth about any degree, on any day, in one place, for free.

04/05

Years of Research

100%

Independent · No Ads

Founder & Vision

✓

Integrity promises

What we will never do

We will not publish unverified claims.

If we cannot back a detail with at least two independent sources or direct faculty confirmation, it does not appear on a ThuSh LK degree page. Ever.

We will not hide our corrections.

When we are wrong, we say so — in public, with the date of the correction and a clear note on what changed. A hidden fix is a broken trust.

We will not sell placement or visibility.

No university, institute, or business can pay to appear higher or better on ThuSh LK. The platform is funded entirely by sales of our own branded scientific calculators.

We will not keep old data frozen.

University curricula change. Our data changes with them. Every page is stamped with its last-verified date so students always know how fresh the information is.

We will not charge students — ever.

Every core tool — Z-Score, Degree Finder, Comparison, Cut-Off tracker, AI risk, career explorer — is free and will remain free to every Sri Lankan A/L and university student.

⚠ Open Correction

Found something that looks wrong?

If any detail on ThuSh LK contradicts your handbook, your faculty, or your own experience — tell us. Every confirmed correction is published publicly with the date of the fix.

Submit a Correction

Data Sources ThuSh LK

Every data point, verified at the source.

Six pillars of verified truth.

Turning raw sources into trusted content.

Automation risk, measured for every degree.

The tools we built — and how.

The thinking behind the platform.

Found something that looks wrong?