*4.2. Principal Component Analysis*

Pearson correlation matrices and PCA construed the datasets. Principal components were generated using varimax rotation, and this yielded variables that contribute more and other variables that contribute less. Multi-variate analysis was used in the PCA to transform a significant set of correlated variables into a minor set of uncorrelated variables. The interrelationships among the variables can be highlighted using covariance by this tool, and it is also called a dimensionless reduction tool. We can use PCA to know the associated chemicals construed as variable loadings on certain groundwater quality factors. Two significant eigenvalues, i.e., PC1 and PC2, were observed in the 40 groundwater samples with 18 parameters, constituting 35 and 12% of the variance. PC3 exhibited 10% of variance but PC4 and PC5 exhibited variance less than 10%. The first five components exhibited eigenvalues that are greater than 0.5. It is assumed that the factor loading value near +/− 1 exhibits a strong correlation. If the value if greater than 0.5, it is significant. PC1 exhibits 35% of variance about significant loadings of K, Pb, Al, As, TDS, nitrate, sulfate, and pH, and is shown in Table 3.


**Table 3.** Principal Component Loadings.

PC2 showed 12% variance associated with significant loadings of Ba, HCO3, and pH. The component characteristics and component loadings for this data are presented in Tables 3 and 4. PC2 is loaded with Ba, HCO3, and pH, whereas PC3 is loaded with Cd, Cr, and Mn. PC4 is loaded with Zn, Al, and pH and PC5 is loaded with Cu and Ni. Temperature, Cr, Ni, and Mn exhibited higher uniqueness values, suggesting that these variables have limited commonality. As evidenced by systematic data analysis, PC1 exhibited significant cations and anions due to anthropogenic and natural sources. *Water* **2022**, *14*, x FOR PEER REVIEW 10 of 20 **Table 4.** Component characteristics. **Variable Model Nugget Ratio (%)**  Cd S \* 0


**Table 4.** Component characteristics. Cr P 63.456 Cu C 34.148

S \* = stable model.

The cos values are employed to know the representation's quality, and the individuals closer to the center of the plot are assumed of limited or low importance for the reported first components. The low cos values are shown in blue and high cos values in red (Figure 4). The contribution of the variables with sample points is presented in Figure 5. The cos values are employed to know the representation's quality, and the individuals closer to the center of the plot are assumed of limited or low importance for the reported first components. The low cos values are shown in blue and high cos values in red (Figure 4). The contribution of the variables with sample points is presented in Figure 5.

**Figure 4.** PCA for variables. **Figure 4.** PCA for variables.

**Figure 5.** Biplot with sample points and variables. **Figure 5.** Biplot with sample points and variables.
