THE PENGUIN PARADOX
This week’s dataset involved deep diving into penguin habits to understand their behavior.
The Palmer Penguins is a dataset constructed by Dr. Kirsten Gorman and relates to the structural size measurements of 3 species of penguins: adult, male and female Adelie, Chinstrap and Gentoo penguins. The data was collected at Palmer station Antarctica LTER between the period 2007–09. For each species, 4 structural size measurements were collected: bill length, bill depth, flipper length, and body mass. In total 344 samples were collected (however 2 samples have missing structural size measurements).
The data on the 3 different species of Penguins was collected from 3 islands in the Palmer archipelago in Antarctica. These islands are the Dream island, Torgerson island and Biscoe island.
What worked for the visualization?
1. A scatter plot depicting the relationship between 2 structural size measurements is ideal.
2. The labels are accurate and easy to read.
3. The differentiation in color across marks based on species helps in understanding how each species correlated between the 2 measurements.
What did not work for the visualization:
1. The segregation of species based on color would have been enough. Creating different shapes for different species is unnecessary and adds to visual clutter.
2. The prevalence of gridlines across the scatter plot is not needed. It does not enhance the analysis.
3. Showing variation across the 3 structural measurements by species would have made the analysis more insightful and thought provoking.
My interpretation:
1. I wanted to understand the correlation between the 3 structural size measurements of Adelie, Gentoo and Chinstrap penguins using scatter plot analysis.
2. A basic understanding of the data gave me insight into the characteristics of the penguin species. Gentoos are the largest of the Palmer penguins. Male Gentoos have an average body mass of 5485 gms with an average flipper length of 221.54 mm. The female Gentoos are 17% less heavy than their male counterparts but their flipper lengths are nearly as long as that of the males with an average length of 213 mm.
3. Chinstrap and Gentoo penguins have the longest bills (culmen length). Male culmen lengths average 51mm and 49 mm respectively while female culmen lengths average 46mm and 45 mm respectively.
4. Adelie and Chinstrap penguins have the deepest bills (culmen depth). Male culmen depths for both species average 19mm. Culmen depths for females of both species average 18 mm.
5. There is a positive correlation between body mass and flipper lengths across all species of penguins.
6. There is a positive correlation between body mass and bill lengths across Gentoo and Adelie penguins. The same cannot be said about Chinstrap penguins.
7. There is no such relation between body mass and bill depths across all species of penguins. The correlation reverses in this case implying that the species exhibit what is called ‘The Simpson Paradox’.
Simpson’s paradox, or the Yule–Simpson effect, is a phenomenon in probability and statistics, in which a trend appears in several different groups of data but disappears or reverses when these groups are combined. It is sometimes given the descriptive title ‘reversal paradox’.
