IoT Sensor Data vs. Open Data Platforms: A Study on Weather Nowcasting Accuracy

Egbulam, Henry Ugochukwu (2025) IoT Sensor Data vs. Open Data Platforms: A Study on Weather Nowcasting Accuracy. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (906kB) | Preview
Preview	PDF (Configuration Manual) Download (388kB) | Preview

Abstract

Accurate hyper-local weather nowcasting is essential for applications ranging from smart home automation systems to outdoor activity planning. However, the reliance on popular open data platforms often fails to capture site-specific microclimatic variations. This study investigates what is termed as the "nowcasting edge" gained by integrating data from a low-cost, personal IoT weather station with standard API data. In order to test this, a robust data pipeline was implemented to synchronise high-frequency sensor readings of temperature, humidity, and pressure collected via a Raspberry Pi-based weather station with data polled from the OpenWeatherMap API. A Stacked GRU neural network was chosen as the optimal architecture for this time-series prediction task. A comparative experiment was conducted between a model trained with IoT-enhanced data and a baseline model relying solely on API data. The results produced confirm the hypothesis to be true. The IoT-enhanced model significantly outperformed the baseline API only model. The RMSE was reduced for temperature by 79.21% and pressure by 90.99% at the 5-minute horizon. While the advantage of the IoT-enhanced model reduces over the forecast window, it still retained a significant 30-83% performance lead at the 60-minute mark. These findings demonstrate that hyper-local sensor data is a critical component for high-accuracy weather nowcasting.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Hasanuzzaman, Mohammed UNSPECIFIED
Subjects:	G Geography. Anthropology. Recreation > GE Environmental Sciences T Technology > TK Electrical engineering. Electronics. Nuclear engineering > Telecommunications > Computer networks > Internet of things G Geography. Anthropology. Recreation > GE Environmental Sciences > Earth sciences > Atmospheric science > Meteorology > Weather
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	07 Sep 2026 10:25
Last Modified:	07 Sep 2026 10:35
URI:	https://norma.ncirl.ie/id/eprint/9856

Actions (login required)

View Item