Key Concepts Learned + Notes
- Importance of Data Visualization and Communication
- Example of Anscombe’s Quartet: Charts with the same summary statistics, but the way the data is represented would suggest otherwise.
- Poor policy decisions can result from misunderstood data. Recall the Jurjevich et al. study that found 72% of Portland Census tracts having unreliable child poverty estimates. This uncertainty was never communicated by planners.
- A note that block groups, since they’re smaller, can have high margins of error! Imagine (an example) just two people getting surveyed in a block group.
- Grammar of Graphics Principles (ggplot2 philosophy)
- Data
- Aesthetics: aes(); connection between data variables and visual properties of the plot
- Geometries: geom_points(); how to display the data - points, lines, etc.
- Visual: Adding scales, themes, facets, annotations, etc.
- Exploratory Data Analysis (EDA)
- What does the data looking like? (distributions, missing values)
- What patterns exists? (relationships, clusters, trends)
- What’s unusual? (outliers, anomalies, data quality issues)
- What questions does this raise? (hypotheses)
- Reliability and data quality?
Coding Techniques/Technical Notes
- ggplot2
- ggplot(data = yourdata) + aes(x = variable1, y = variable2) + geom_something() + additional_layers()
- Again, aesthetics are used to map data to visual properties. You shouldn’t be setting themes here (for example), but if you want the color to change based on a variable, you can do it here!