	INTAKE OF 16:0	SATURATED FATTY ACID INTAKE	CHOLESTEROL INTAKE	INTAKE OF 18:0	MONOUNSATURATED FATTY ACID INTAKE	TOTAL n9 MONOUNSATURATED FATTY ACID INTAKE	SPICE INTAKE
INTAKE OF 16:0	1.000000	0.985032	0.653803	0.955669	0.978646	0.974084	-0.108781
SATURATED FATTY ACID INTAKE	0.985032	1.000000	0.675034	0.985912	0.941295	0.932656	-0.115341
CHOLESTEROL INTAKE	0.653803	0.675034	1.000000	0.645144	0.652127	0.638503	0.100037
INTAKE OF 18:0	0.955669	0.985912	0.645144	1.000000	0.903516	0.894436	-0.109789
MONOUNSATURATED FATTY ACID INTAKE	0.978646	0.941295	0.652127	0.903516	1.000000	0.999485	-0.067324
TOTAL n9 MONOUNSATURATED FATTY ACID INTAKE	0.974084	0.932656	0.638503	0.894436	0.999485	1.000000	-0.069250
SPICE INTAKE	-0.108781	-0.115341	0.100037	-0.109789	-0.067324	-0.069250	1.000000

Looks like our MUFA and SFA intake is corrlated too strongly amongst themselves to look at these variables other than as one. Lets look at some more stuff lower down the line: ```python interesting = chinastudy89[["SATURATED FATTY ACID INTAKE", "MONOUNSATURATED FATTY ACID INTAKE", "CHOLESTEROL INTAKE", "ANIMAL PROTEIN INTAKE", "MEAT INTAKE"]] interesting.corr() ```

	SATURATED FATTY ACID INTAKE	MONOUNSATURATED FATTY ACID INTAKE	CHOLESTEROL INTAKE	ANIMAL PROTEIN INTAKE	MEAT INTAKE
SATURATED FATTY ACID INTAKE	1.000000	0.941295	0.675034	0.634446	0.803585
MONOUNSATURATED FATTY ACID INTAKE	0.941295	1.000000	0.652127	0.576671	0.724846
CHOLESTEROL INTAKE	0.675034	0.652127	1.000000	0.917797	0.800527
ANIMAL PROTEIN INTAKE	0.634446	0.576671	0.917797	1.000000	0.815691
MEAT INTAKE	0.803585	0.724846	0.800527	0.815691	1.000000

The correlations speak for themselves, the top X correlations for longevity are relatively strongly correlated amongst themselves. It is likely that multiple of these columns are indeed major contributors to longevity, yet given the inter-food correlations, it is quite impossible to isolate these from the variables that are just along for the ride. So from this data, can we claim that either SFA, cholesterol, meat or animal protein contributes to increased longevity? No, that would be a way too strong claim based on this data. We can say though that it is *very unlikely* any of the top X variables, including SFA, cholesterol, animal protein and meat has significant *detremental* effect on longevity. A claim that while the book does not make it explicitly is implied by the long list of adverse health affects the book claims flow causally from the consumption of these predominently animal-sourced substances. The above analysis is trivially simple. Just data preprocessing simple correlations and filtering, using the most high-level overview the data allows for regarding food intake and health. The possible effect of different foods on longevity. No massaging of the data, no exposure to birthday paradox spuriousness, no statistical tricks, questionable adjustments for, etc. Just a simple matrix of correlation cooeficients showing just how unlikely the sum of all claims made by Campbell must be considered to be. Campbell managed to spin this study to convey a proteinophobic and anti animal-sourced foods message that, as could be expected, has been rather well received by the vegan community. So much so that that community is now setting out to get Campbell nominated for a Nobel prize. I'll let Campbell's data vs his book speak for themselves here. Does Campbell deserve a being nominated for such a prestigious prize? Well, I'll leave you all with this thought. ![download.jpeg](https://cdn.steemitimages.com/DQmTeRPN8oq2r31kctoZs3C3qycQSS8MTo1xHqvKmkWoGYq/download.jpeg) The Jupyter Notebook of this blog post is available [here](https://rmeijer.home.xs4all.nl/ChinaStudyDemo.ipynb). Please use is to look at the data and if there is any oversight in this almost trivial analysis please comment below.