Data Viz Tips & Tricks
Eric Ekholm
CCPS
About Me
About These Materials
- All plots created using R and the
ggplot2visualization library - Slides created using Quarto
- Data is publicly available from VDOE (& is also available in this repository).
- All code/slides/data available in this Github repository
Types of Data Viz
- Generally, there are 2 “types” of data viz we do:
- Embedded in reports/slides (usually static)
- Standalone (usually interactive; e.g. a dashboard)
- This presentation is focused on the first (although some principles may apply to the second as well)
Framing
- Imagine you’re asked to conduct an analysis of your division’s historical performance on the grades 3-8 reading SOLs.
Overarching Principle: Use Data to Tell a Story
- Embrace the fact that visualizations are subjective
- Make your viz decisions shape the story you want to tell
- Tailor the story you’re telling to the decisions you’re trying to inform
Tip #1: Tell Viewers the Takeaway
Generic Title
Opinionated Title
Discussion
- What are your approaches?
- When might you not want to use this?
- Other thoughts/questions?
Tip #2: Avoid Extraneous Information
Too Much
Just Enough
Discussion
- What are your approaches?
- When might you not want to use this?
- Other thoughts/questions?
Tip #3: Use Aesthetics that Match Your Story
Compare to Avg
Compare to Baseline
Compare to Other
Compare to Other
Discussion
- What are your approaches?
- When might you not want to use this?
- Other thoughts/questions?
Tip #4: Build to Complex Visuals
- Generally, try to keep visualizations simple and aligned with your key takeaways
- If you need to do something more complex, build to it over several plots
Start With
Then
Finally
Discussion
- What are your approaches?
- When might you not want to use this?
- Other thoughts/questions?
Wrapping Up
- Remember that you are telling a story with your data
- Decisions you make about how to present your data should align with your story
- Less is usually more
- If you need a complex visual, build to it