Project 2

Project Objective:

As a Data Scientist at an international humanitarian NGO, the task is to strategically allocate $120 million in newly raised funding. The core objective is to identify and prioritize countries in dire need of aid based on socio-economic and health indicators that reflect overall development levels. Through data-driven insights, we aim to ensure that funds are directed towards countries where they will have the most substantial impact.

Data Preparation and Analysis:

We begin by gathering socio-economic and health data for a diverse set of countries. Key indicators such as GDP per capita, income levels, life expectancy, health expenditure, and other socio-economic factors are selected as they effectively reflect the well-being and development needs of each country. Outliers within these indicators are identified and treated, as clustering algorithms are sensitive to extreme values.

Methodology:

1. Principal Component Analysis (PCA):

   • To reduce the dimensionality of the data, we apply PCA, allowing us to focus on the most influential features that capture variance in development indicators. This helps to simplify the data and improves the clarity of clustering results.

2. Clustering Analysis:

   • K-Means Clustering: Using K-Means clustering, we classify countries into distinct groups based on similar socio-economic and health characteristics. The optimal number of clusters is determined, and countries within each cluster are analyzed for common traits.
   • Hierarchical Clustering: To validate our findings and gain additional insights, we apply hierarchical clustering. This method provides a visual dendrogram, which helps in identifying the natural groupings of countries and understanding the relationships between clusters.

Results and Recommendations:

The clustering analysis reveals distinct socio-economic levels among countries, classifying them into groups such as:

• High-income, high-development countries
• Moderate-income countries with mid-level socio-economic indicators
• Low-income countries facing significant health and socio-economic challenges

By comparing the characteristics of each cluster, we identify the countries in the low-income, high-need group as priority candidates for humanitarian aid. These countries demonstrate the most urgent need for assistance based on their socio-economic and health indicators.

Conclusion:

Using PCA and clustering techniques, we provide the NGO's leadership with a clear, data-backed recommendation to focus resources on countries where the humanitarian impact of aid will be the highest. This strategic allocation ensures that the $120 million funding addresses the most pressing needs, supporting sustainable improvements in health, education, and living standards for the world’s most vulnerable populations.