In a regression, the residual is what's left after the model has said everything it can. That's the part I like: take a headline number, find the public data behind it, and see what it leaves out.
I'm a final-year Statistics student at UiTM, and most of what's here is built on Malaysian open data: OpenDOSM, the Ministry of Higher Education, MetMalaysia and the Budget documents.
Looking for a Data Analyst internship in Kuala Lumpur, available now until January 2027. LinkedIn · Portfolio
Budget 2027 announced a RM2,500 wage floor for graduates. Half of 2024's first-degree graduates were already at or below it, but that half is not spread evenly.
The numbers behind the chart are in data/, and the
full analysis is in fresh-grad-salary-model.
| The question | What the data said | Built with |
|---|---|---|
| How many graduates does a RM2,500 wage floor reach, and what does it cost? | 49.7% earned RM2,500 or less in 2024. In simulation the graduate wage bill rises 7.1% to 14.4%, depending on who is covered. fresh-grad-salary-model | Python, statsmodels, Excel |
| What does a graduate hire cost an employer at the new wage benchmarks? | Employer cost and take-home pay at RM2,000, RM2,500 and RM3,400 a month, in a formula-driven workbook. belanjawan-2027-wage-cost-model | Excel, Python |
| Which months of Malaysian trade don't fit the pattern? | 320 months of OpenDOSM trade data, with 8 export anomaly months and 25 sector spikes flagged. dosm_trade_engine | Python, pandas, Power BI |
| How much do grocery prices really differ by store format? | A 16-staple basket is 14% cheaper in hypermarkets (RM110.96) than wet markets (RM126.58), but within-format shop dispersion dominates. 20M+ PriceCatcher rows modeled with two-way fixed effects. pricecatcher-retail-pricing | DuckDB, Python, Power BI |
| How far is the market fuel price from the subsidised one this week? | A live tracker of the gap for RON95 and diesel, refreshed twice a week. budi-pulse · live | Python, PostgreSQL, Streamlit |
| Who leaves a telco, and when? | 26.54% churn across 7,043 customers; month-to-month contracts are 88.55% of those who left. DSC651-TelcoChurn-PowerBI | Power BI |
| Can a model catch fraud that is 0.17% of transactions? | A detection pipeline for a heavily imbalanced dataset. audit-fraud-detection | Python, SQL |
| What does the Budget actually say? | Answers drawn from the official Budget 2026 documents, with page-level citations. belanjawan-bot | Python, RAG |
Also here: perasaan-malaysia, a scheduled pipeline for MetMalaysia's 7-day forecasts; malaysia-inclusive-economy-blueprint · live, an interactive fiscal policy simulator; esg-multiagent-auditor, an event-driven Scope 1-3 GHG auditing architecture; and wealthwatch, a personal finance tracker that is live.
Analysis Python (pandas, statsmodels), R, SQL (PostgreSQL, T-SQL), SAS, SPSS
Reporting Power BI (DAX, star schemas), Excel, Tableau, Looker Studio, Streamlit
Methods multiple linear regression, hypothesis testing, regression diagnostics, anomaly detection
Model summary: luqman
======================================================================
Dep. variable: answers you can check
Method: least squares
Sample: Oct 2023 - Feb 2027, B.Sc. (Hons.) Statistics, UiTM
----------------------------------------------------------------------
Notes:
[1] Every number on this page comes from a repository linked above.
[2] Simulated data is labelled as simulated.
[3] The header and chart are drawn by scripts/build_assets.py.
======================================================================