The Effectiveness of Hong Kong's Typhoon Warning Policy: Public Perception and Adaptive Behavior (Quantitative questionnaires and qualitative web crawler data)

Project Title: 

The Effectiveness of Hong Kong's Typhoon Warning Policy: 

Public Perception and Adaptive Behavior

Description:This dataset was collected for a Final Year Project (FYP) at Hong Kong Baptist University. It aims to investigate public sentiment and reactions towards typhoon warning signals in Hong Kong.

The dataset includes:

Quantitative Survey Data (n=350): Anonymized responses from a questionnaire.

The main collection target is young people in Hong Kong, a small number of middle-aged and elderly people (excluding minors).

The following content is a reverse scoring question, and reverse scoring has been processed

I believe the Observatory needs to issue Typhoon Signal No. 8 warning earlier.

I think the business and school closure warning issued one day in advance for a super typhoon (Typhoon Signal No. 10) does not leave me enough time to prepare for holiday-related matters (such as finishing work/study, purchasing supplies, etc.).

I think the business and school closure warning issued one day in advance for a super typhoon (Typhoon Signal No. 10) has disrupted my work or study plans.

I think the business and school closure warning issued one day in advance for a super typhoon (Typhoon Signal No. 10) has had a negative impact on my personal travel or social arrangements.

If the weather is sunny with no wind or rain when the one-day advance warning is issued, it would cause me to underestimate the actual intensity of the approaching super typhoon (e.g., Ragas)

Qualitative Scraping Data (4,000+ entries): Social media comments crawled from Facebook related to typhoon Ragasa (2025).

Facebook data is more representative of the opinions of middle-aged and elderly citizens compared to Threads. Therefore, Facebook data primarily collects information on the perceptions of elderly people regarding typhoon response measures.

Cantonese Sentiment Lexicon: A specialized coding table for Cantonese expressions and their sentiment scores used in this study. Based on school requirements, the paper could only be completed by one person alone, so codes were not cross-validated by a second researchers

Analysis Matrices: Co-occurrence matrices and word frequency data derived from the collected text.

     Notes:

All personal identifiable information (PII) has been removed to ensure participant privacy.

In accordance with internet research ethics, the original account links and key identifying information of the speakers have been removed from the dataset used in this study. Third-party usernames mentioned in the comments are for contextual reference only and are not intended to identify any specific individual.

This dataset contains raw social media comments, which may include profanity or offensive language. These are preserved for the integrity of linguistic and sentiment analysis.

The questionnaire content has been approved by the Hong Kong Baptist University's ethics committee.