Environmental Sensor and ML Experiment
An experiment collecting temperature and humidity data to test models of air-conditioning use.
Built a sensor-to-SQL workflow and tested statistical and predictive methods for air-conditioning decisions.
From raw inputs to working systems.
I build the connections between collecting data and putting it to use: sensor experiments, database pipelines, review tools and automated reports. These projects trace that work from the input to the interface.
Selected Systems
Collection / Processing / Review
Records and Exception Monitoring
A shared review interface for recurring records, scheduled checks and potential gaps across separate workflows.
Brought missing records and potential gaps into one place for review and follow-up.
Record Matching and Verification
A review dashboard combining event logs, OCR-aware fuzzy matching and structured investigation of inconsistent identifiers.
Made repeated records and recognition mismatches easier to investigate in a structured review view.
Browser-Based Workflow Capture
A Chrome extension connecting structured checklist entry with an existing reporting workflow.
Simplified checklist entry while keeping the existing reporting workflow in place.
Automated HDB Trends Reporting
A monthly reporting workflow combining HDB resale transactions, town-level geography, and generated summaries.
Produced repeatable price comparisons and town-level summaries with less manual report assembly.
Technical Case Studies
Data Pipelines and Experiments
Recurring Report Automation
A Python and SQL pipeline that standardised inputs from several sources and generated recurring reporting outputs.
Reduced manual report assembly and made recurring outputs more consistent.
Decision-Support Dashboards
A shared reporting interface bringing external indicators and internal performance measures into one review workflow.
Brought recurring external and internal indicators into a shared reporting interface.
Home SQL Server for Sensor Inputs
A Raspberry Pi server I built with Apache, PHP, and MariaDB to collect sensor readings for analysis.
Set up a self-hosted database that received sensor readings and made them available for analysis.
Daily Stock Data Ingestion Pipeline
An R workflow that collected daily stock data, joined company information, and wrote derived features to SQL for dashboards.
Created a repeatable daily feed of market data and derived features for SQL-backed dashboards.
Live Feedback Pipeline From Google Forms to SQL
A scheduled R pipeline moving new Google Form responses into SQL Server for Power BI reporting.
Kept feedback dashboards updated within minutes of submission without paid middleware.
Sensor Uptime Monitoring Dashboard
A dashboard checking sensor posting times, missing readings, and delays in a SQL data feed.
Helped identify late or unstable sensors before their missing data affected analysis.
These cases describe the methods and workflows I built. Operational projects are presented in general terms; client information, proprietary data, and confidential implementation details are omitted.