NATURAL LANGUAGE PROCESSING EMPLOYING SENTIMENT ANALYSIS ON THE PUBLIC VOICE OF FILIPINOS DURING CRISIS SITUATIONS

Abstract

Aims: This Natural Language Processing study examined the sentiments of Filipinos in crisis situations using Twitter data. The goal was to identify the prevalent emotional tone, categorize sentiments, and explore variations across different types of crises.

Study design: Qualitative Research using Natural Language Processing

Place and Duration of Study: Philippines, 2019 - 2023

Methodology: There were 10,000 tweets that narrate the public voice of Filipinos during crisis situations that were analyzed using sentiment analysis. They were mined using Apify Easy Twitter Search Scraper. Results were analyzed using Orange Software and Plutchik’s Emotional Wheel via Box Plot.

Results: Results show that the dominant public voice is neutral, this is followed by positive and negative tones. This means that Filipinos mainly share information, describe their situations, and seek clarifications during crises. It is further revealed that joy was the most dominant emotion, followed by sadness, trust, surprise, fear, disgust, anticipation, and anger. This means that even in crisis, Filipinos wish people’s safety, pray for the situation and give jokes to lighten the mood. The order of dominant emotions or tone for earthquake, inflation, and typhoon followed a pattern of neutral, positive and negative. The pandemic stood out, featuring a dominant neutral tone, followed by negative and then positive.

Conclusion: These Findings Offer Valuable Insights for Refining Government Crisis Response Strategies. Additionally, This Underscores the Significance Of Recognizing And Understanding Emotions In Communication, Providing Valuable Guidance For Educators And Students.