Truth About Nutritional Information in Recipes

So much has been said about Proteins, Fat & Calories in recent years, that the single biggest challenge faced when trying to answer the question is how to “separate the wheat from the chaff.” Protein and fats are what our body uses and they all have calorie counts. Too many and we get fat, too few and we starve. In this article, we look at the truth about nutritional information in recipes on the Epicurious website.

The dataset is available with over 20k recipes. In the dataset calories, protein, fat and sodium have almost same missing values. Let us explore these:


Looking at some of the NaN columns such as Scotch Cooler, Fish Brine and others, it looks like all have the same rows. so we will drop these rows.

Also, we will explore where the ratings is zero.

title       1296
rating      1296
calories    1296
protein     1296
fat         1296
sodium      1296
There are 1296 with zero ratings. 
The recipes have very high quantity of sodium and very high calories, mostly receive zero ratings.

The rating seems to be aggregated as it has 8 unique values:

[ 2.5 4.375 3.75 5. 3.125 1.875 0. 1.25 ]

The recipes with 4.375 have higher calories when compared to best rating 5. So recipes with more calories get’s rating around 4 than getting 5.

The above plot just proves that there is a positive correlation between fat and calories.

Most of the 5 point rating in the dataset are low in fat, low in calories, high in proteins and low in sodium. There are 135 recipes with sodium above the mean of 6252. However, upon further exploration, it seems that the most of the sodium is guesswork in the recipes. Or the number of portions in the recipes or the measurement will help to find more about the sodium.

Will there be any correlation between Nutrition?

List the numerical features decendingly by their correlation with Protein:

sodium: 	0.7492865810748386
calories: 	0.7428158481088419
fat: 	0.7121938623158276
rating: 	0.013970568366160188

Hence, recipes with high protein also have a high amount of sodium, calories and fat. Is this because high-protein foods are very frequently high-fat foods: meat, milk, and cheeses? Also, 1 gramme of protein is equal to 4 calories. Hence, too much of protein leads to excess calories.

If you would like to check out the notebook, please click this link .

Hope you enjoy reading this article. In the case of any queries or suggestions, please feel free to leave a comment or get in touch with me.

Leave a Reply