Intelligent Automation of climate change media sentiment analysis using news APIs, Power Automate, and Power BI
Subramaniyamge, Ruwani (2025)
2025
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2025111227811
https://urn.fi/URN:NBN:fi:amk-2025111227811
Tiivistelmä
The objective of this thesis was to develop and execute an intelligent automated media monitoring system for climate change news coverage based on API integration, Microsoft Power Automate, and Power BI. The research questions focused on how low-code tools are utilized to automate sentiment analysis of climate news, what process can be used to obtain reliable sentiment analysis results, and what kind of methods enable efficient data collection and visualization.
The research was practical and based on a constructive research method combined with an agile development model. The theoretical framework identified four core interrelated areas, such as media monitoring theory, automation technologies, APIs to pull data and integrate sources of data, and business intelligence presentations. The three-tier architecture included acquisition, processing, and presentational data. Data reliability and system performance were assessed both technically and analytically, and the result indicated that the process worked without programming skills.
The research result showed that low-code automation process is effective for media monitoring and sentiment analysis. The developed system enables real-time monitoring of sentiment and trends in news about climate change, which improved the possibilities of utilizing information in decision-making. Based on results, it can be recommended that similar solutions be utilized in environmental organizations, media agencies, and research institutions that need objective and up-to-date monitoring.
The limitations of the system were found related to the interface search limitations, a narrow selection of sources, language dependencies, and challenges of sentiment analysis in understanding scientific articles, which affect the comprehensiveness of the data. In the future, the system will be developed by using premium API services, expanding the number of sources, increasing multilingualism, and improving the accuracy of sentiment analysis with specialized machine learning models. Finally, the research shows that already existing low-code tools can achieve significant results in successfully automating media monitoring.
The research was practical and based on a constructive research method combined with an agile development model. The theoretical framework identified four core interrelated areas, such as media monitoring theory, automation technologies, APIs to pull data and integrate sources of data, and business intelligence presentations. The three-tier architecture included acquisition, processing, and presentational data. Data reliability and system performance were assessed both technically and analytically, and the result indicated that the process worked without programming skills.
The research result showed that low-code automation process is effective for media monitoring and sentiment analysis. The developed system enables real-time monitoring of sentiment and trends in news about climate change, which improved the possibilities of utilizing information in decision-making. Based on results, it can be recommended that similar solutions be utilized in environmental organizations, media agencies, and research institutions that need objective and up-to-date monitoring.
The limitations of the system were found related to the interface search limitations, a narrow selection of sources, language dependencies, and challenges of sentiment analysis in understanding scientific articles, which affect the comprehensiveness of the data. In the future, the system will be developed by using premium API services, expanding the number of sources, increasing multilingualism, and improving the accuracy of sentiment analysis with specialized machine learning models. Finally, the research shows that already existing low-code tools can achieve significant results in successfully automating media monitoring.