Showing relevant statistics

Concetti di comunicazione dei dati

Hadrien Lacroix

Curriculum Manager

Data storytelling road

Puzzle piece with data word highlighted

Absolute
- Difference between 2018 and 2017 sales
- Absolute change and relative change depend on the quantity

bars showing absolute change of sales in two years

Relative
- Percentage variation 2018 from 2017
- Small numbers more significant than reality

bars showing relative change of sales in two years

Absolute

Graph showing two barplots with no difference

Relative

Graph showing a barplot with one bar higher than another one

Quotient of two variables
- Revenue per customer (total product revenue/number customers)
Normalize values = better comparisons

bars showing revenue per customer for chocolate and chips

bars showing mean

boxplot showing median price for two products

Representative value:
- Totals / counts
- Mean
- Median
Mean can be misleading (outlier)
Distribution of the data
Example:
- 2019 US average salary: $51,916.27
- 2019 US median salary: $34,248.45

What is p-value?

Convention:
- Value less than 0.05: statistical significance
- Values close to 0.05: weak indicator

What is it not?

¹ P-Values: Misunderstood and Misused. Front. Phys. 2016.

Concetti di comunicazione dei dati