An introductory course in Visualisations
“The simple graph has brought more information to the data analyst’s mind than any other device.” — John Tukey
Topics
- Purpose of Data Visualisations: EDAs and the Narrative
- Exploratory Data Analysis (EDA): Visualisations are an essential part of the exploratory analysis, that helps understand the data, such as the distribution of the outcome variable, or potential links between covariates. You need to understand the data before you build a model!
- The research narrative: Summarise visually the results of your research in an easy-to-understand memorable graphic.
- Design principles: Simplicity, the Principle of Proportional Ink (Bergstrom and West):
- What not to do: Bullshitting, misrepresentations, flawed statistical “analysis” (deliberate or incompetent)
- The Mechanics of Data Visualisations: Fundamentals of DataVis and the Grammar of Graphics
- Fine-tuning your graphs: Themes, colour, avoiding overplotting, annotations
- Maps and geo-spatial analysis (including IPUMS, NHGIS, ERA5 APIs)
- Interactive graphics and dynamic Dashboards (bread-and-butter work of the professional Data Scientist) Building a dashboard using shiny.
- Graphical research design: Binscatter plots, event-studies, regression discontinuity design (RDD).