Ship Weather Monitoring System
Built data sanitization tools for Navy ship weather monitoring systems.
Overview
Early in my time at Forward Slope Inc., I built Java tools that cleaned up the weather data coming off sensors on Navy ships, so the fleet could actually trust it when making operational decisions.
The Challenge
Raw sensor data is messy. Readings came in with anomalies and errors, different sensor types needed normalizing against each other, invalid values had to be caught rather than passed downstream, and the cleaned output had to meet strict quality standards.
What I Built
I wrote Java tools for processing the data streams. On the validation side: range checks on temperature, pressure, and humidity; temporal consistency checks; cross-sensor correlation; and anomaly detection for the readings that didn't add up. On top of that came the transformations, including unit conversions and normalization, coordinate-system transformations, time-zone handling for a distributed fleet, and format standardization for the systems downstream. Automated tests, data-quality metrics, and alerts rounded it out so problems surfaced quickly.
Where It Landed
The result was cleaner data for the fleet's weather systems, a lot less manual review, more reliable weather-dependent operations, and a set of reusable processing components that other projects picked up. It was a good early grounding in ETL patterns, data validation, and writing code that has to handle every odd edge case the ocean can throw at it.