Key Concepts Learned
- Visualization matters: summary statistics can obscure patterns; outliers may represent critical communities.
- Ethical responsibility to report MOE, show uncertainty, and avoid misleading visuals.
- Grammar of Graphics (ggplot2):
Data → Aesthetics → Geometries → Visuals, with additional layers like scales, themes, and annotations.
- Exploratory Data Analysis (EDA): a detective process of exploring distributions, relationships, anomalies, and reliability before modeling.
- Data Joins: left, right, inner, and full joins for integrating datasets in policy analysis.
Coding Techniques
- ggplot2 basics:
- Structure:
ggplot(data) + aes(x=var1, y=var2) + geom_*() + layers.
- Aesthetic mappings: x, y, color, fill, size, shape, alpha.
- Adding layers with
+ for customization.
- EDA tools: histograms, scatterplots, boxplots for detecting distributions and outliers.
- Data reliability: use coefficient of variation (CV) thresholds to categorize estimates as reliable, somewhat reliable, or unreliable.
- dplyr joins:
left_join(), right_join(), inner_join(), full_join().
Questions & Challenges
- How to best visualize uncertainty so that policymakers actually understand and act on it.
- Practice needed in customizing ggplot themes for clear and professional presentation.
- Still clarifying when to collapse categories or aggregate geographies to reduce statistical uncertainty.
Connections to Policy
- Poor visualization (e.g., ignoring MOEs) can directly misinform policy and harm vulnerable communities.
- Reporting reliability aligns with the AICP Code of Ethics — transparency is both technical best practice and ethical obligation.
- Visualizations tailored for different audiences (legislators, community groups, media) can shift how data drives decisions.
Reflection
- The most interesting insight was Anscombe’s Quartet: identical summary statistics but radically different patterns when visualized — proving the necessity of graphs.
- I’ll apply this by always starting analysis with EDA and incorporating data quality checks into my visuals.
- This week reinforced that visualization is not just aesthetic — it is a tool for ethical communication in public policy.