I love making visualizations of data. I have no training on this, so there's a high probability that any good matching of colors you see (if any) was suggested by someone else.
(1) Alcohol consumption trajectories in People Living with HIV (PLWH)
Caption: predicted (upper panels) and observed (bottom panels) alcohol consumption trajectories in PLWH that are moderate drinkers at baseline. The entire set of observations is plotted in the bottom panels, but each observation is weighted (through changing transparency) by the probability of membership to each of the 3 classes.
From: Bilal U, et al. Measuring Alcohol Trajectories in People Living with HIV: Modeling Finite Mixtures of Multinomial Distributions. Under Review.
(2) Visualizing n-dimensional correlation networks
Caption: Weighted correlation network of obesogenic characteristic of 600 communities in Pennsylvania. Darker nodes have a higher bivariate correlation with childhood obesity (% children with >85pct of weight for age). Thicker and darker edges indicate a higher bivariate correlation between two nodes.
From: Knapp EA, et al. Obesogenic environments in childhood in Pennsylvania, A network approach. Under Review
(3) 3D Visualizations in 2D through color
Caption: Change in the unemployment/mortality association over time and across levels of change in social protection expenditure. Left panel: magnitude of the association. Right panel: significance (p-value) of the association. Dashed and dotted lines represent the abyss of significance at the 0.05 and 0.01 level, respectively.
From: Bilal U, et al. Economic Growth and Mortality: Do Social Protection Policies Matter? In Press