Women in the Workplace

Interactive scatterplot (plotly and shiny packages)


March 4, 2019

Notable topics: Interactive scatterplot (plotly and shiny packages)

Recorded on: 2019-03-04

Timestamps by: Alex Cookson

View code



Writing a custom function that summarizes variables based on their names (then abandoning the idea)


Using complete.cases function to find observations that have an NA value in any variable

Using subsetting within a summarise function to calculate a weighted mean when dealing with 0 or NA values in some observations

Debugging what is causing NA values to appear in the summarise output (finds the error at 13:25)

Hypothesizing about one sector illustrating a variation of Simpson's Paradox


Creating a scatterplot with a logarithmic scale and using scale_colour_gradient2 function to encode data to point colour


Creating an interactive plot (tooltips show up on hover) using ggplotly function from plotly package

Fiddling with scale_size_continuous function's range argument to specify point size on a scatterplot (which are encoded to total workers)

Explanation of why healthcare sector is a good example of Simpson's Paradox


Starting to create a shiny app with "occupation" as only input (many tweaks in subsequent minutes to make it work)


Tweaking size (height) of graph in shiny app

Summary of screencast