Fig. 2.

Multiple Correspondence Analysis (MCA) based on questionnaire data to generate lifestyle score. A) MCA was applied to categorical questionnaire data (38 manually curated questions; 21 on assets, 11 on food and 6 on housing) (N = 203 individuals). Data points are coloured based on location. Ellipses reflect the data spread at a level of confidence of 95%. Density plots show the distribution of PC1 (lifestyle score) (x-axis) and PC2 (y-axis) score. B) Comparisons of PC1 (lifestyle score) and PC2 across locations. Global significance was assessed using analysis of variance (ANOVA) and post hoc tests between locations were performed using Tukey HSD tests. Asterisks denote statistical significance (NS, non-significant; ∗, p ≤ 0.05; ∗∗, p ≤ 0.01; ∗∗∗, p ≤ 0.001, p ≤ 0.0001). C) Coordinates of each variable category (a.-t.; see E) across dimensions 1 and 2. Variable categories with similar profiles are grouped together. D) Cumulative contributions (in percentage) of the variable categories by questionnaire data category (i.e. housing, assets and food). E) Contributions (in percentage) of variable categories to PC1 or lifestyle score. Bars are coloured based on whether a variable was associated with a high (>zero) or low (<zero) lifestyle score.