Page 154 - FULL REPORT 30012024

6.2 Limitations of Project

This research encountered specific data-related challenges that impacted the

effectiveness and contextual relevance of both the dashboard and the machine
learning model. The dashboard, a critical tool for visualizing stroke mortality

trends, utilized data that was only updated up to the year 2021. This limitation
restricts the dashboard's ability to provide current insights, which are

essential for accurately reflecting ongoing and emerging stroke-related health

trends. Figure 6.1 in the study illustrates this limitation, as highlighted by a
user during the acceptance testing phase.

Figure 6.1 Screenshot of a user feedback from the user acceptance testing

Furthermore, the machine learning model employed in this study utilized a
dataset sourced from Kaggle. While Kaggle datasets are renowned for their

reliability, the dataset used may not accurately reflect the specific

demographic and health profile peculiar to the Malaysian population. For a
model that aims to predict stroke risk in Malaysia, a dataset derived from

local sources or one closely mirroring the demographics and health
characteristics of Malaysians would be more appropriate. This would

enhance the model's precision and relevance in predicting stroke risks within
the Malaysian context.

The limitations in data sourcing underscore the necessity for access to more

recent, localized, and demographically representative data. Such data is

crucial for improving the effectiveness and accuracy of both the dashboard
and the machine learning model, particularly in addressing stroke-related

health issues within Malaysia. The need for targeted data underscores the

137