INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue X October 2025

www.rsisinternational.org

Page 9516

A Data-Driven Approach to Flood Risk Assessment and Public Sentiment

Analysis

Md. Shadman Zoha

, Nor Aiza Moketar

, Massila Kamalrudin

, Suriati Akmal

, Noorrezam Yusop

Mohd Riduan Ahmad

, Ariff Idris

, Takeshi Morimoto

Centre of Technology for Disaster Risk Reduction (CDR), Faculty of Information, Communication and

Technology, University Technical Malaysia Melaka, Melaka, Malaysia

Centre of Technology for Disaster Risk Reduction (CDR), Institute of Technology Management and

Entrepreneurship, University Technical Malaysia Melaka, Melaka, Malaysia

Centre of Technology for Disaster Risk Reduction (CDR), Faculty of Electronics and Computer

Technology and Engineering, University Technical Malaysia Melaka, Melaka, Malaysia

Faculty of Science and Engineering, Kindai University Kowakae, Higashiosaka City, Osaka, Japan

DOI: https://dx.doi.org/10.47772/IJRISS.2025.910000779

Received: 07 November 2025; Accepted: 14 November 2025; Published: 24 November 2025

ABSTRACT

Flooding remains as the most frequent and destructive natural disaster impacting Malaysia which causing

significant disruptions across social, economic and environmental systems. This study addresses the need to

integrate physical risk assessment with public sentiment analysis to strengthen disaster management. Historical

flood records from 1967 to 2023 were analyzed together with flood-related news articles, utilizing

geographical risk mapping and transformer-based as well as keyword-driven sentiment analysis. The results

identified Kelantan and Terengganu as the highest-risk states and revealed dominant emotions of fear and

frustration in media coverage, with 74.5% of articles emphasizing rescue operations and only 11.8% focusing

on recovery. These findings highlight critical gaps in long-term flood resilience communication and planning.

By integrating data-driven flood risk assessment with sentiment insights, the study offers a more

comprehensive understanding of flood impacts, supporting more targeted disaster preparedness,

communication strategies and policy development in Malaysia.

Keywords – Data-Driven, Decision-making, Flood, Risk Assessment, Sentiment Analysis

INTRODUCTION

Floods are among the most frequent and destructive natural disasters globally, affecting millions of people

every year and causing significant economic, environmental and social losses (Saharudin et al., 2023). In

Malaysia, floods occurred almost every year, primarily during the monsoon season (Rosmadi et al., 2023). In

recent decades, the occurrence of flood events in Malaysia has shown an increasing pattern, caused by rapid

urbanization, land-use changes and growing impacts of climate change (Abd Majid et al., 2020)(Romali &

Yusop, 2021). The shape of the land and weather conditions also play a huge role in causing flood in Malaysia

(Elsheikh et al., 2015). The impact of flooding is countless, ranging from damage to critical infrastructure,

disruption of daily life, transportation, essential services and economic activities (Saad et al., 2024)(Kumar &

Jha, 2023) (Aminah Shakirah et al., 2016). As a result, flood risk management has become a national priority,

demanding more proactive and integrated strategies.

Traditional flood risk assessments mostly focused on physical and environmental dimensions such as

topography, rainfall intensity and hydrological modelling which often utilizing geospatial and multi-criteria

decision-making methods to map hazard zones and assess vulnerabilities. These approaches have produced

valuable insights for mitigation and infrastructure planning. Moreover, the existing flood management effort in

Malaysia, guided by official framework such as Directive 20, primarily emphasized operational coordination

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue X October 2025

www.rsisinternational.org

Page 9517

during disaster event (Rosmadi et al., 2025). However, these strategies often overlook the social and emotional

responses of the affected communities. Recent studies emphasized that understanding the public perception,

sentiment and emotional reaction is essential for improving early warning communication, disaster

preparedness and community resilience (Kaklauskas et al., 2024). In Malaysia, studies integrating public

sentiment into flood risk frameworks are still limited, leaving a significant gap in the country’s capacity to

understand and respond to the numerous impacts of floods.

This study seeks to address this gap by combining geospatial flood risk assessment with sentiment analysis of

flood-related news articles. Specifically, this research is guided by the following research question: How is the

integrated geospatial flood risk assessment with sentiment analysis enhance understanding of flood impacts

and support more effective disaster management in Malaysia?

The motivation for this research arises from the need to bridge technical hazard models with human centred

insights. Understanding how communities perceive and emotionally respond to flood occurrences provides

decision-makers a more robust evidence bases for developing targeted communication strategies, enhancing

coordination, and building trust between authorities and the public. Furthermore, combining sentiment data

with risk mapping enables more precise resource allocation and better prioritization of high-risk areas,

resulting in increased readiness and resilience. By examining historical flood data from 1967 to 2023 together

with the analysed public sentiment extracted from flood-related news articles, this research offers new

perspective into how floods are experienced, perceived, and represented at the societal level.

Related Works

The current flood risk assessment generally incorporates hazard, exposure and vulnerability layers within

Geographical Information Systems (GIS) and multicriteria decision analysis (MCDA) to provide actionable

risks maps. Studies conducted in Malaysia and similar contexts shows that integrating topographical features

(elevation, slope), hydrometeorological factors (rainfall, flow accumulation), land-use and socio-economic

vulnerability enables the identification and classification of areas with varying risk levels. This approach

allows authorities and planners to prioritise high-risk zones for mitigation measures, infrastructure

reinforcement and flood shelter planning (Usman Kaoje et al., 2021). Similarly, the study in the Kanjiro River

Basin, Indonesia (Hatta et al., 2025) demonstrates how the systematic integration of hazard, exposure and

vulnerability can improve the precision and efficacy of local-scale flood risk assessment. These systematic

approaches enhance decision-making, improve resource allocation and facilitate long-term flood resilience

planning.

Recent years have seen a growing interest in applying sentiment analysis as a complementary tool for disaster

management, particularly during flood events. In the study by U. H. Hair Zaki, R. Ibrahim, S. Abd Halim, and

T. Yokoi (Hair Zaki et al., 2017), the authors investigated how sentiment analysis can help to manage the flood

disasters. This study highlighted the critical importance of understanding the public opinion during such

events, especially by gathering significant information from the large number of social media posts. The study

emphasized that spotting posts showing fear, concern or panic could play a key role in responding to disasters.

Although sentiment analysis is often used for things like product reviews, there hasn’t been much research on

using it during natural disasters, especially in the Malay language, which is mainly spoken in Malaysia. This

gap highlighted the need for more localized and linguistically relevant approaches to disaster sentiment

monitoring in Malaysia.

In similar study, the authors introduced a structured process model for analyzing flood-related sentiments on

social media (Zaki et al., 2017). They combined machine learning technique along with the Rational Unified

Process (RUP) and Service-Oriented Modeling and Architecture (SOMA) frameworks to develop an approach

to classify and interpret the public emotions during the flood events. This method is useful for the current

research as it demonstrates how integrating process models with sentiment analysis can provide structured,

scalable insights for flood disasters response and coordination.

The study by M. A. Saddam, E. K. Dewantara, and A. Solichin (Saddam et al., 2023) applied sentiment

analysis to evaluate the public perceptions of flood management in Jakarta, Indonesia. The authors used the

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue X October 2025

www.rsisinternational.org

Page 9518

Support Vector Machine (SVM) to classify the tweets into positive, neutral and negative categories. The study

showed strong results, with the 88.6% accuracy, 88.6% precision and 89.4% recall which demonstrates the

potential of machine learning techniques in capturing and interpreting public sentiment during disaster events.

The findings suggest that similar methodologies could be adapted to the Malaysian context to provide timely,

data-driven insights to community reactions during floods.

These studies demonstrate how sentiment analysis can be strategically incorporated into disaster management

systems to enhance traditional environmental and logistical data. However, most of the existing research

focuses on either generic applications or non-Malay linguistic contexts, leading to a clear research gap in

localized sentiment analysis for flood disaster management in Malaysia. Addressing this gap can enhance

situational awareness, improve community engagement, and facilitate more efficient emergency response

strategies.

METHODOLOGY

This study used a clear step-by-step method to analyse the flood risks and emotions from news about floods.

The methodology has four main steps: Data Collection, Pre-processing, Analysis and Result as shown in the

Figure 1.

Figure 1. Step-by-step research flow

Data Collection

There were two main data sources used in this study: (1) historical flood data and (2) flood-related news

articles. The historical data dataset was downloaded from the following link: https://www.emdat.be/. This

dataset includes detailed records of flood events, including dates, locations and impacts shown in Figure 2. The

dataset loaded using the Pandas library, a widely used tool for data analysis in Python.

Figure 2. Historical flood data

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue X October 2025

www.rsisinternational.org

Page 9519

Flood-related news articles collected from the FloodList website at https://floodlist.com/tag/malaysia to gather

relevant articles. The data was collected through web scraping using the BeautifulSoup and Requests libraries

in Python. The collected data is stored in a list and converted into a pandas DataFrame for analysis. This

dataset includes title, articles text, link and date which will be analysed for sentiment.

Preprocessing

Geocoding and State Extraction

Geocoding and state extraction were used to identify the location of flood events. For historical data, reverse

geocoding converted latitude and longitude into state names using the Google Maps API. For news articles, a

predefined list of Malaysian states was used to extract state names from article titles.

Data Cleaning and Preparation

Missing values into the dataset were replaced with zeros using the fillna(0) method. Numeric columns like

'Start Year' and 'End Year' were converted to numeric format for proper data handling. The dataset was also

filtered for the articles to include only flood events by selecting rows where the 'Disaster Type' column

contained the keyword 'Flood'.

Analysis

Risk Score Calculation

Risk Score calculated based on the historical flood data. The score integrates key factors, including the Total

Affected Population, Total Deaths, and the Consumer Price Index (CPI). Using this approach, flood-prone

areas are classified into three risk levels:

 High Risk: Severe flood impact with high affected population, deaths and economic losses (CPI-

adjusted).

 Medium-High Risk: Significant flood impact with moderately high affected population, noticeable

deaths and economic losses.

 Medium Risk: Moderate impact with manageable population, deaths and losses.

 Low Risk: Minimal impact with low affected population, few/no deaths and minor losses.

Sentiment Categorization

In this paper, we used a combination of transformer-based sentiment analysis and a rule-based keyword

matching technique to classify and analyse the news articles about floods in Malaysia. The sentiment analysis

was performed using Hugging Face's transformers library, specifically the pre-trained BERT (Bidirectional

Encoder Representations from Transformers) model, which excels in natural language processing (NLP) tasks.

This model processes the text by splitting it into manageable chunks, evaluates the sentiment (positive,

negative) for each chunk, and aggregates the results to provide an overall sentiment score.

Sentiment Analysis

To refine sentiment analysis, keyword-based classification was used across several categories. Severity was

determined by keywords related to "loss_of_life", "property_damage" and "logistical_issues". Emotional

impact was classified using words linked to fear, frustration, sadness and helplessness. Keywords like

"rescued", "donation" and "aid" identified articles on rescue and recovery efforts. Neutral reports, including

weather updates, government actions, community support and health advisories, were also categorized. All

these keyword groups are shown in Figure 3.

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue X October 2025

www.rsisinternational.org

Page 9520

Figure 3. Sentiment Categorization

RESULT AND DISCUSSION

Historical Data Analysis

The analysis showed the distribution of floods across different states in Malaysia. Figure 4 presented a bar

chart highlighting the states that experienced the most flood events from 1967 to 2023. Figure 5 displayed a

time series chart showing the number of floods each year, helping to identify trends over time. This

visualization highlighted years with severe floods and years with fewer incidents. Understanding these patterns

helped improve flood management and preparedness strategies.

Figure 4. Flood occurances by states

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue X October 2025

www.rsisinternational.org

Page 9521

Figure 5. Flood frequency over time

The risk levels for the state were categorized based on the number of flood occurrence count. The visualization

showed in Figure 6, that some states, particularly those with more severe events, were marked with higher risk

levels.

Figure 6. Flood risk level by state in Malaysia

The classification of flood events by subtype provided a detailed view of flood occurrences. Figure 7 showed

the yearly distribution of different flood subtypes from 1967 to 2023, helping to track how often each type

occurred over time. Figure8 compared the three main flood subtypes: general floods, riverine floods and flash

floods. Categorizing floods this way gave better insight into their causes, characteristics and impacts on

different regions. This classification helped improve flood risk assessment and mitigation planning.

Figure 7. Flood risk level in Malaysia by state

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue X October 2025

www.rsisinternational.org

Page 9522

Figure 8. Flood subtype in Malaysia

A comprehensive risk map was created, visually displaying in Figure 9, the relative flood risk of each place.

This map helps to identify high-risk areas based on historical flood data.

Figure 9. Flood prone areas

Using a dataset of flood incidents, categorized and visualized the top 10 river basins based on the frequency of

flood occurrences presented in Figure 10. This visualization helps in understanding which river are at the

greatest risk. The use of horizontal bar charts allows for a clear comparison of flood frequency across these

basins, highlighting critical hotspots of flood activity.

Figure 10. Top ten flood incidents by river basin

Overall, this section provided a quantitative insights and visualization, that how floods affect in Malaysia. It is

important to find out the areas that get flooded often in Malaysia.

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue X October 2025

www.rsisinternational.org

Page 9523

Sentiment Analysis

This section presents the sentiment analysis from the flood-related news articles. Figure 11 shows the negative

and positive emotions in the news where the occurrences of negative sentiment is higher than the positive . It

provides understanding about how strong these emotions were about floods. Figure 12 presents the overall

distribution of emotions, a compressive overview of the people’s emotions in the news. By studying these

patterns, we can understand how the public felt and reacted to the flood events over time. This analysis is

important to understanding public awareness, response and concern during the flood events.

Figure 11. Sentiment Distribution

Figure 12. Sample of positive and negative keywords

Figure 13 shows how many news articles specifically talked about the rescue operations, recovery efforts or

both. By looking at this data, we can see how often the flood responses are reported and highlighted. This

helps us to understand the trends in news coverage and the focus on rescue and recovery over time. It also

helps to evaluate how well the public is informed about flood management actions.

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue X October 2025

www.rsisinternational.org

Page 9524

Figure 13. Rescue and recovery mentioned in articles

In this section, the articles are sorted by how serious the flood impacts are. It covers things like problems with

logistics, loss of life, damage to property or no serious damage at all. Figure 14 illustrates how frequently each

of these categories is mentioned in the news articles, providing a clearer picture of the media's focus on

different aspects of flood severity. This breakdown aids in understanding the narrative of flood coverage and

how often the severity of different impacts is underrepresented or emphasized. The emotional impact reflected

in the flood-related articles by categorizing them into four key emotional responses: no emotion, frustration,

fear, and sadness. Figure 15 visualizes the frequency of each emotional impact, highlighting how the media

portrays the emotional response to floods in Malaysia.

This result shows how people reacted emotionally and socially during flooding events. They were frustrated

with slow recovery, scared of rising the water level and sad about the loss and damage. These feelings help us

to understand what they went through into the time of flood.

Figure 14. Severity sentiment distribution

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue X October 2025

www.rsisinternational.org

Page 9525

Figure 15. Emotion impact distribution

Figure 16 shows how often neutral topics like community support and weather reports appear in the news

articles about floods in Malaysia. Articles about community support highlight how people and groups come

together to help those affected, showing strength and unity during tough times. On the other hand, weather

reports give clear, factual updates about the conditions, which help keep the public informed and support better

decision-making during flood events. By analyzing these neutral reporting categories, we can understand how

the media balances emotional narratives with factual updates, thereby influencing public perception and

response during the flood emergencies. Figure 22 the heatmap showed the link between the severity in flood

news and the emotions expressed. Most articles, even when reporting serious issues like property damage or

loss of life, did not show any strong emotion. However, frustration appeared more often when discussing

logistical issues and loss of life. Fear and sadness were rarely used. This meant that while most reports stayed

neutral, some emotions did come through, especially in more serious or stressful situations. This helped us

understand how the media presented emotional impact during flood events.

Figure 16. Neutral reporting categories

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue X October 2025

www.rsisinternational.org

Page 9526

Figure 17. Severity vs emotional impact

The number of negative sentiment articles about floods in different states of Malaysia presented into the Figure

18. It highlighted which states had the more negative coverage in the media. States with the higher negative

sentiment may have faced more serious flood impacts or had issues with response efforts.

Figure 18. Historical flood data

DISCUSSION

Spatial Distribution of Flood Risk

The geospatial analysis demonstrates that Kelantan and Terengganu continue to be Malaysia's most flood-

prone states, with persistently high flood frequencies over the last five decades. These findings are consistent

with prior hydrological and GIS-based evaluations, which have identified the east coast's vulnerability to

monsoon-driven riverine floods. The historical trends identified in this study show that flood occurrences have

become more frequent and severe, which is consistent with broader regional patterns associated with

urbanisation and climate variability. The spatial concentration of flood events in specific river basins

highlights the ongoing vulnerability of critical infrastructure and settlements, with substantial implications for

disaster preparedness, infrastructure investment, and long-term resilience planning. By establishing a robust

spatial baseline, this study provides a critical foundation for integrating physical risk with social dimensions of

disaster management.

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue X October 2025

www.rsisinternational.org

Page 9527

Public Sentiment and Disaster Communication

The analysis of sentiment in flood-related news articles indicated that fear and frustration were the

predominant emotions expressed during flood events. Emotional responses serve as indicators not only of

disaster severity but also of public perceptions regarding institutional readiness and the effectiveness of

response mechanisms. The significant focus on rescue operations (74.5%) relative to recovery narratives

(11.8%) indicates a structural communication imbalance that favours immediate emergency responses over

long-term community recovery efforts. This pattern corresponds with findings from disaster communication

research indicating that emergency-focused narratives can increase public anxiety and limit awareness of

resilience measures. The prevalence of neutral reporting indicates that Malaysian media frequently portray

floods as normal events, which may lead to public desensitisation. The integration of sentiment assessment

into flood management systems may enhance adaptive communication strategies, improving community trust

and compliance in response efforts.

Regional Variation and Policy Implications

The state-level sentiment analysis indicates significant regional variations, with higher negative sentiment

primarily observed in developed states such as Negeri Sembilan and Kuala Lumpur. This may indicate

increased exposure, more extensive media coverage, or perceived deficiencies in local response capabilities.

The findings indicate that flood communication strategies must avoid a single approach. They need to be

customised to the socio-economic and infrastructural context of each region. High-risk rural areas may benefit

from enhanced early warning systems and community-based preparedness programs, while urban regions

might need improved inter-agency coordination and information transparency. The insights correspond with

the overarching policy framework of Malaysia’s National Security Council Directive No. 20, which prioritises

localised disaster response. Integrating sentiment data into spatial risk mapping enables policymakers to

prioritise resources more effectively, enhance targeted messaging, and strengthen community engagement.

Limitation, Contribution and Future Research

This study opens to several limitations. First, sentiment extracted from news articles may inadequately

represent the complexity of public perception, as media coverage often reflects editorial decision and language

framing. Second, the temporal resolution of the data may not correspond precisely with flood event timelines,

which creating potential biases in emotion classification. Third, while the BERT-based model demonstrated

robust performance, incorporating additional local language models could further improve sentiment detection

accuracy. Despite these limitations, this study contributes methodologically by integrating geospatial and

sentiment analysis to provide a richer, multidimensional understanding of flood impacts in Malaysia. Future

work could expand this framework by embedding the causal relationships between flood severity indicators

and public emotion into disaster communication simulations and decision-support dashboards to build dynamic

and localized flood intelligence systems. This will enable agencies to anticipate public reactions, optimize

communication strategies, and strengthen trust during crises. Together, these advancements will extend the

current study beyond analytical insight toward actionable, impact-oriented disaster management solutions.

CONCLUSION

This study provides a comprehensive understanding of flood risks in Malaysia by integrating historical flood

risk analysis with sentiment analysis derived from news articles. The findings highlight the spatial distribution

of flood-prone areas and reveal how communities have experienced and responded to the flood events over the

years. The analysis indicates that certain states particularly Kelantan and Terengganu have faced recurrent

flood incidents, indicating higher levels risk and the needs for sustained mitigation and preparedness planning.

The sentiment analysis further reveals strong public emotions, including fear and frustration, in conjunction

with accounts of rescue and recovery efforts. These dual perspectives capture not only physical and

infrastructural impacts of floods but also the social and emotional dimensions, providing more comprehensive

understanding of disaster impact in Malaysia. A key insight from this study is that 74.5% of flood-related news

articles emphasized on rescue operations, while only 11.8% highlighted recovery and long-term resilience

efforts. The difference indicates an essential communication gap in how flood events are presented to the

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue X October 2025

www.rsisinternational.org

Page 9528

public. When media narratives focus predominantly on emergency response, communities may perceive floods

as short-term crises rather than recurring hazards that ongoing preparedness, resilience-building and post-

disaster rehabilitation. To address this communication imbalance, this study recommends introducing a post-

disaster communication framework to ensure more balanced and informative flood reporting. The framework

should require media coverage to highlight not only rescue operation but also post-disaster recovery and long-

term resilience initiatives. This should provide unified communication guidelines and strengthen coordination

between media organizations and disaster authorities to improve public awareness and support more effective

community preparedness.

The flood risk map is useful for the disaster management agencies, facilitating more targeted response

strategies and more efficient resource allocation. By combining the data-driven risk assessment with sentiment

analysis, this research contributes to better flood preparedness, response planning and overall resilience in

Malaysia. Future work may focus on developing a flood prediction model based on the identified flood-prone

areas, which could enhance early warning systems and strengthen disaster management strategies in Malaysia.

ACKNOWLEDGMENT

The study is funded by SATREPS grant (GERAN ANTARABANGSA - SATREPS/2023/FKEKK/ A00052)

in collaboration between Japan Science and Technology Agency (JST, JPMJSA2210) and Japan International

Cooperation Agency (JICA) and Ministry of Higher Education (MOHE) of Malaysia by FRGS grant

(FRGS/1/2022/FTMK/F00525) and. It is also promoted by MOHE in Malaysia. This work is additionally

supported by the Kesidang Scholarship.

REFERENCES

1. Abd Majid, N., Rizal Razman, M., Zarina Syed Zakaria, S., Farid Ahmed, M., & Azwani Zulkafli, S.

(2020). Flood disaster in Malaysia: approach review, causes and application of geographic information

system (GIS) for Mapping of flood risk area. In Copyright@ EM International.

2. Aminah Shakirah, J., Sidek, L. M., Hidayah, B., Nazirul, M. Z., Jajarmizadeh, M., Ros, F. C., &

Roseli, Z. A. (2016). A Review on Flood Events for Kelantan River Watershed in Malaysia for Last

Decade (2001-2010). IOP Conference Series: Earth and Environmental Science, 32(1).

https://doi.org/10.1088/1755-1315/32/1/012070

3. Elsheikh, R. F. A., Ouerghi, S., & Elhag, A. R. (2015). Flood Risk Map Based on GIS, and Multi

Criteria Techniques (Case Study Terengganu Malaysia). Journal of Geographic Information System,

07(04), 348–357. https://doi.org/10.4236/jgis.2015.74027

4. Hair Zaki, U. H., Ibrahim, R., Abd Halim, S., & Yokoi, T. (2017). A review on service oriented

architecture approach in flood disaster management framework for sentiment analysis: Malaysia

context. Frontiers in Artificial Intelligence and Applications, 297, 362–377.

https://doi.org/10.3233/978-1-61499-800-6-362

5. Hatta, M. P., Pongmanda, S., Suprapti, A., Sari, K., Rijal, S., Samad, W., & Fadlin, F. (2025). The

Spatial Model of Flood Risk in the Kanjiro River Basin, North Luwu District, Indonesia. Engineering,

Technology and Applied Science Research, 15(3), 22848–22856. https://doi.org/10.48084/etasr.10235

6. Kaklauskas, A., Rajib, S., Piaseckiene, G., Kaklauskiene, L., Sepliakovas, J., Lepkova, N.,

Abaravicius, Z., Milevicius, V., Kildiene, S., & Sapurov, M. (2024). Multiple criteria and statistical

sentiment analysis on flooding. Scientific Reports, 14(1), 1–16. https://doi.org/10.1038/s41598-024-

81562-0

7. Kumar, N., & Jha, R. (2023). GIS-based Flood Risk Mapping: The Case Study of Kosi River Basin,

Bihar, India. Engineering, Technology and Applied Science Research, 13(1).

https://doi.org/10.48084/etasr.5377

8. Romali, N. S., & Yusop, Z. (2021). Flood damage and risk assessment for urban area in Malaysia.

Hydrology Research, 52(1), 142–159. https://doi.org/10.2166/NH.2020.121

9. Rosmadi, H. S., Ahmed, M. F., Mokhtar, M. Bin, & Lim, C. K. (2023). Reviewing Challenges of Flood

Risk Management in Malaysia. In Water (Switzerland) (Vol. 15, Issue 13). Multidisciplinary Digital

Publishing Institute (MDPI). https://doi.org/10.3390/w15132390

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue X October 2025

www.rsisinternational.org

Page 9529

10. Rosmadi, H. S., Ahmed, M. F., Radwan, N., Mokhtar, M. Bin, Lim, C. K., Halder, B., Scholz, M.,

Alshehri, F., & Pande, C. B. (2025). Flood Management Framework for Local Government at Shah

Alam, Malaysia. Water (Switzerland), 17(4). https://doi.org/10.3390/w17040513

11. Saad, M. S. H., Ali, M. I., Razi, P. Z., & Ramli, N. I. (2024). Flood Risk Management in Development

Projects: A Review of Malaysian Perspective within the Sendai Framework for Disaster Risk

Reduction 2015-2030. CONSTRUCTION, 4(2), 103–117.

https://doi.org/10.15282/construction.v4i2.10592

12. Saddam, M. A., Dewantara, E. K., & Solichin, A. (2023). Sentiment Analysis of Flood Disaster

Management in Jakarta on Twitter Using Support Vector Machines. Sinkron, 8(1), 470–479.

https://doi.org/10.33395/sinkron.v8i1.12063

13. Saharudin, M. A. I. Bin, Rosli, M. A. N. Bin, Handayani, D. O. D., Basri, A. B. B., Attarbashi, Z. S., &

Suryady, Z. (2023). Flood Forecasting Using Weather Parameters. 2023 IEEE 9th International

Conference on Computing, Engineering and Design, ICCED 2023.

https://doi.org/10.1109/ICCED60214.2023.10425318

14. Usman Kaoje, I., Abdul Rahman, M. Z., Idris, N. H., Razak, K. A., Wan Mohd Rani, W. N. M., Tam,

T. H., & Mohd Salleh, M. R. (2021). Physical flood vulnerability assessment using geospatial

indicator‐based approach and participatory analytical hierarchy process: A case study in Kota Bharu,

Malaysia. Water (Switzerland), 13(13). https://doi.org/10.3390/w13131786

15. Zaki, U. H. H., Ibrahim, R., Halim, S. A., Khaidzir, K. A. M., & Yokoi, T. (2017). Sentiflood: Process

model for flood disaster sentiment analysis. 2017 IEEE Conference on Big Data and Analytics, ICBDA

2017, 2018-January, 37–42. https://doi.org/10.1109/ICBDAA.2017.8284104