Data Science Desktop Survival Guide
by Graham Williams
Faceted Location Scatter Plot
ggplot(aes(x=date, y=max_temp)) +
geom_point(alpha=0.05, shape=".") +
geom_smooth(method="gam", formula=y~s(x, bs="cs")) +
theme(axis.text.x=element_text(angle=45, hjust=1)) +
Partitioning the dataset by a categoric variable reduces the blob effect for big data. The plot uses location as the faceted variable to separately plot each location's maximum temperature over time. Notice the seasonal effect across all plots, some with quite different patterns.
The plot uses facet_wrap() to separately plot each location. Using ggplot2::geom_point() with alpha= reduces the effect of overlaid points as does using smaller dots on the plots by way of shape=. Together this works to de-clutter the plot and improves the presentation with an emphasis on the patterns. The x-axis tick labels are rotated using angle=45 within ggplot2::element_text() to avoid the labels overlapping. The hjust=1 forces the labels to be right justified.