Data & Visualizations
Why do we create data visualizations?
to see patterns and answer questions
What is metadata?
data about data
Chart showing how many times each value appears
— bar chart
What does filtering do?
— selects subset of data
What is training data?
— data used to teach AI
What warning reminds us not to assume one variable causes another?
correlation does not equal causation
Name one thing metadata helps with
organizing/finding/managing data
Chart grouping numeric data into buckets
— histogram
When should data be cleaned?
— incomplete or inconsistent data
What happens if training data is biased?
— biased results
Visualizations help reveal patterns invisible when looking only at what?
a data table
Can metadata change without affecting primary data?
yes
Chart showing combinations of two columns
— crosstab
If only female legislators are needed
— filter
Why judging appearance with AI is problematic?
— unintended bias
What type of statement is 'What does the data show?'
fact
Give one example of messy data
inconsistent spelling/abbreviations/capitalization
Best chart for spotting numeric trends
— scatter plot
Which comes first: clean/filter or visualize? —
clean/filter
What is algorithmic bias?
— unfair outcomes from data/design
What step focuses on finding patterns?
visualize and find patterns
Goal of cleaning data
— fix inconsistencies without changing meaning
When is a bar chart not useful?
— when values are mostly unique
Why might charts differ across students?
— messy or differently filtered data
One way to reduce ML bias
— diverse training data/human oversight