Suyog Dharmadhikari & Magaritte Nguyen 12 October, 2022
This document is a vignette to show how to retrieve data from an API. To demonstrate, we will be interacting with the Spoonacular Recipe and Food API. We are going to to build a few functions to interact with some of the endpoints and explore some of the data that we can retrieve for the Spoonacular API.
We will use a function that connects with complexSearch, which is an
end point of the API search. It returns a list of lists containing
recipe names and other variables of interest that have default values or
can be user defined. The user may either enter the type of cuisine,
recipes with a specific ingredient, number of recipes to output based on
modifications provided, type of diet, and protein range. We have another
function that can retrieve the data for multiple cuisines specified by
the user as well. We have a third function which can also be used to
contact other endpoint of the API, findByNutrients to get all the
recipes based on carbohydrate content.
To use the functions for interacting with the Spoonacular Recipe and Food API, we used the following packages:
tidyverse: tons of useful features for data manipulation and visualizationjsonlite: API interaction
In addition to those packages, I used the following packages in the rest of the document:
Note: User needs to go to
API_link, create a
profile and generate a key. This key will be used to authenticate
the user for each API calls.
For this section, we have written some functions that will connect the user to the Spoonacular API and returns a object with the required information specified by the user.
In this function we call complexSearch endpoint of the Spoonacular
API. The user can use this function to get the data from the API for
different types of cuisine, recipes with a specific ingredient, maximum
number of recipes to output, type of diet, and protein range. The
function is user friendly and handles upper and lower case scenarios
with spaces. GET from httr is used to connect with the API and
retrieved the required information while fromJSON is used to convert
the raw data into a readable format.
Just some extra information on how to use the my_spoonacular_api: we
are only able to specify one cuisine at a time, or else a message might
appear letting the user know what is appropriate in this field. We can
also specify one ingredient, but if there is nothing that matches, then
the results will just be 0 recipes.if the user would like some nutrient
information as well, they can input TRUE or FALSE for this parameter.
Lastly, we have different diets that the user can specify, such as
vegetarian etc., and the amount of protein in grams (min and max range).
NOTE: user cannot fetch more than 100 records in the single call as it is restricted by the API.
# Query will only return words in the title of the recipe
# If set to true, you get nutritional information about each recipes returned.
# 100 is the max number of results you can get back for a single query.
my_Spoonacular_API <- function (cuisinetxt ="American", ingredient = "fruit", nutrientinfo = "TRUE", number=100, diet = "", minProtein = 10, maxProtein = 100 ){
#valid cuisines to enter into the cuisine parameter
#using tolower() to make everything lowercase for east of search
#using str_replace_all for searches that are more than one word e.g. Latin American
if ( ! any(str_replace_all(tolower(cuisinetxt), " ", "-") %in% str_replace_all( tolower(c("African", "American", "British", "Cajun", "Caribbean", "Chinese", "Eastern European", "European", "French", "German", "Greek", "Indian", "Irish", "Italian", "Japanese", "Jewish", "Korean", "Latin American", "Mediterranean", "Mexican", "Middle Eastern", "Nordic", "Southern", "Spanish", "Thai", "Vietnamese")), " ", "-" ) ) ) { stop ('Please only include an acceptable cuisine input seen below:
"African", "American", "British", "Cajun", "Caribbean", "Chinese", "Eastern European", "European", "French", "German", "Greek", "Indian", "Irish", "Italian", "Japanese", "Jewish", "Korean", "Latin American", "Mediterranean", "Mexican", "Middle Eastern", "Nordic", "Southern", "Spanish", "Thai", "Vietnamese"
*** If you include more than one cuisine seperated by a comma, ignore this message.')
} #end of cuisine check
#using GET to return info about a topic that we are interested in
myGET <- GET(url=paste0("https://api.spoonacular.com/recipes/complexSearch?cuisine=", cuisinetxt, "&includeIngredients=", ingredient, "&addRecipeNutrition=", nutrientinfo, "&number=",number,"&diet=",diet,"&minProtein",minProtein,"&maxProtein",maxProtein,"&apiKey=",params$`api-key`))
#looking at content (raw data) and converting it into something we can read from myGET
#data
spoonacular <- fromJSON(rawToChar(myGET$content))
spoonacular$cuisinetxt <- cuisinetxt
#results returned as a dataframe
return(spoonacular)
} #end of function that called APIIn this function we call complexSearch endpoint of the API. It is
similar to my_Spoonacular_API but in this function the user can enter
multiple cuisines to get the combined data of all the cuisines. The user
has to use paste with collapse = "," as the parameter to create a
list of cuisines before passing it to the function.
NOTE: user cannot fetch more than 100 records in the single call as it is restricted by the API.
# Function for multiple cuisines in the cuisinetxt parameter
my_multiCuisine_API <- function (cuisinetxt =" ", ingredient = "", nutrientinfo = "TRUE", number=100, diet = "", minProtein = 10, maxProtein = 100 ){
myGETTemp <- GET(url=paste0("https://api.spoonacular.com/recipes/complexSearch?cuisine=", cuisinetxt, "&includeIngredients=", ingredient, "&addRecipeNutrition=", nutrientinfo, "&number=", number,"&diet=",diet,"&minProtein",minProtein,"&maxProtein",maxProtein,"&apiKey=",params$`api-key`))
return(myGETTemp)
}The findByNutrients endpoint of the API is called in the method below.
The function can assist users in retrieving data on the nutritional
value of recipes. To choose dishes from the specified carbohydrate
content and certain number of recipes output, the user can alter the
carbohydrates range. It returns the object along with some information
on the recipe, its nutrient content, and other parameters. GET
from httr is used to connect with the API and get the required
information while fromJSON is used to convert the raw data into
readable format.
NOTE: user cannot fetch more than 100 records in the single call as it is restricted by the API.
# Looking for the given number of recipes with minCarbs and maxCarbs
findRecipeByNutrients <- function(minCarbs = 10, maxCarbs = 100, numberOfRecipes = 100){
getURL <- paste0("https://api.spoonacular.com/recipes/findByNutrients?minCarbs=",minCarbs,"&maxCarbs=",maxCarbs,"&number=",numberOfRecipes,"&apiKey=",params$`api-key`)
foodData <- GET(getURL)
findByNutrientsDF <- fromJSON(rawToChar(foodData$content))
return(findByNutrientsDF)
}In the code chunk below, we have called my_Spoonacular_API() and
findRecipeByNutrients(), which creates objects for different cuisines
with different ingredients and other parameters. The user can reference
this section of the vignette as an example call, if they want to get
information about different cuisines with different modifications.
#Example calls for my_Spoonacular_API
AmericanCuisineInfo <- my_Spoonacular_API (cuisinetxt ="American", ingredient = "fruit", nutrientinfo = "TRUE", number=100, diet = "", minProtein = 0, maxProtein = 100)
LatinAmericanCuisineInfo <- my_Spoonacular_API (cuisinetxt = "Latin American", ingredient = "salt", nutrientinfo = "TRUE", number=100, diet = "", minProtein = 10, maxProtein = 100)
ChineseCuisineInfo <- my_Spoonacular_API (cuisinetxt = "Chinese", ingredient = "cabbage", nutrientinfo = "TRUE", number=100, diet = "glutenFree", minProtein = 10, maxProtein = 100)
#Example calls for findRecipeByNutrients
dispDF <- findRecipeByNutrients(numberOfRecipes = 20)These are the data frames containing nutrient information about the particular cuisine, which can be used further for our data exploration (EDA).
AmericanCuisineInfotbl <- as_tibble(AmericanCuisineInfo$results)
AmericanCuisineInfotbl## # A tibble: 100 × 35
## vegetarian vegan glutenF…¹ dairy…² veryH…³ cheap veryP…⁴ susta…⁵
## <lgl> <lgl> <lgl> <lgl> <lgl> <lgl> <lgl> <lgl>
## 1 TRUE TRUE FALSE TRUE FALSE FALSE FALSE FALSE
## 2 FALSE FALSE TRUE FALSE TRUE FALSE FALSE FALSE
## 3 FALSE FALSE TRUE FALSE FALSE FALSE FALSE FALSE
## 4 FALSE FALSE FALSE TRUE TRUE FALSE FALSE FALSE
## 5 FALSE FALSE TRUE TRUE FALSE FALSE FALSE FALSE
## 6 FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE
## 7 FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE
## 8 FALSE FALSE FALSE TRUE FALSE FALSE FALSE FALSE
## 9 FALSE FALSE TRUE FALSE FALSE FALSE FALSE FALSE
## 10 FALSE FALSE FALSE TRUE FALSE FALSE FALSE FALSE
## # … with 90 more rows, 27 more variables: lowFodmap <lgl>,
## # weightWatcherSmartPoints <int>, gaps <chr>,
## # preparationMinutes <int>, cookingMinutes <int>,
## # aggregateLikes <int>, healthScore <int>, creditsText <chr>,
## # sourceName <chr>, pricePerServing <dbl>, id <int>,
## # title <chr>, readyInMinutes <int>, servings <int>,
## # sourceUrl <chr>, image <chr>, imageType <chr>, …
LatinAmericanCuisineInfoTbl <- as_tibble(LatinAmericanCuisineInfo$results)
LatinAmericanCuisineInfoTbl## # A tibble: 12 × 35
## vegetarian vegan glutenF…¹ dairy…² veryH…³ cheap veryP…⁴ susta…⁵
## <lgl> <lgl> <lgl> <lgl> <lgl> <lgl> <lgl> <lgl>
## 1 FALSE FALSE TRUE TRUE TRUE FALSE FALSE FALSE
## 2 FALSE FALSE TRUE TRUE TRUE FALSE FALSE FALSE
## 3 FALSE FALSE TRUE TRUE FALSE FALSE FALSE FALSE
## 4 FALSE FALSE TRUE TRUE FALSE FALSE FALSE FALSE
## 5 FALSE FALSE FALSE TRUE FALSE FALSE FALSE FALSE
## 6 FALSE FALSE TRUE TRUE FALSE FALSE FALSE FALSE
## 7 FALSE FALSE TRUE TRUE FALSE FALSE FALSE FALSE
## 8 FALSE FALSE TRUE TRUE FALSE FALSE FALSE FALSE
## 9 TRUE TRUE TRUE TRUE FALSE FALSE FALSE FALSE
## 10 TRUE FALSE TRUE FALSE FALSE FALSE FALSE FALSE
## 11 FALSE FALSE TRUE FALSE FALSE FALSE FALSE FALSE
## 12 FALSE FALSE TRUE TRUE FALSE FALSE FALSE FALSE
## # … with 27 more variables: lowFodmap <lgl>,
## # weightWatcherSmartPoints <int>, gaps <chr>,
## # preparationMinutes <int>, cookingMinutes <int>,
## # aggregateLikes <int>, healthScore <int>, creditsText <chr>,
## # sourceName <chr>, pricePerServing <dbl>, id <int>,
## # title <chr>, readyInMinutes <int>, servings <int>,
## # sourceUrl <chr>, image <chr>, imageType <chr>, …
ChineseCuisineInfoTbl <- as_tibble(ChineseCuisineInfo$results)
ChineseCuisineInfoTbl## # A tibble: 4 × 34
## vegetarian vegan glutenFree dairy…¹ veryH…² cheap veryP…³ susta…⁴
## <lgl> <lgl> <lgl> <lgl> <lgl> <lgl> <lgl> <lgl>
## 1 FALSE FALSE TRUE TRUE TRUE FALSE FALSE FALSE
## 2 FALSE FALSE FALSE TRUE FALSE FALSE FALSE FALSE
## 3 FALSE FALSE FALSE TRUE FALSE FALSE FALSE FALSE
## 4 FALSE FALSE TRUE TRUE FALSE FALSE FALSE FALSE
## # … with 26 more variables: lowFodmap <lgl>,
## # weightWatcherSmartPoints <int>, gaps <chr>,
## # preparationMinutes <int>, cookingMinutes <int>,
## # aggregateLikes <int>, healthScore <int>, creditsText <chr>,
## # license <chr>, sourceName <chr>, pricePerServing <dbl>,
## # id <int>, title <chr>, readyInMinutes <int>, servings <int>,
## # sourceUrl <chr>, image <chr>, imageType <chr>, …
dispDFTbl <- as_tibble(dispDF)
dispDFTbl## # A tibble: 20 × 8
## id title image image…¹ calor…² protein fat carbs
## <int> <chr> <chr> <chr> <int> <chr> <chr> <chr>
## 1 634010 Banana Bread … http… jpg 211 4g 9g 31g
## 2 634141 Banana Oatmea… http… jpg 148 4g 3g 27g
## 3 634931 Best Soft Gin… http… jpg 75 1g 2g 13g
## 4 635856 Braised Pork:… http… jpg 583 16g 47g 31g
## 5 637440 Chapchae (Kor… http… jpg 397 5g 15g 64g
## 6 638496 Chickpea and … http… jpg 272 9g 12g 36g
## 7 643244 Four-Ingredie… http… jpg 162 5g 1g 37g
## 8 649048 Korean Extra … http… jpg 937 56g 51g 60g
## 9 649314 Lavash http… jpg 278 7g 9g 46g
## 10 655249 Peanut Butter… http… jpg 359 8g 23g 34g
## 11 657095 Prawn Curry http… jpg 458 29g 17g 47g
## 12 657226 Pulled Pork S… http… jpg 873 83g 34g 59g
## 13 657889 Raspberry Pea… http… jpg 164 2g 5g 29g
## 14 658151 Refreshing Zu… http… jpg 331 4g 29g 17g
## 15 660525 Soft-Baked Pr… http… jpg 376 7g 22g 41g
## 16 661843 Strawberry Ma… http… jpg 441 3g 25g 55g
## 17 663824 Trinidadian C… http… jpg 554 35g 37g 20g
## 18 716195 Spicy Indian-… http… jpg 134 5g 6g 15g
## 19 1062883 How to Make E… http… jpg 345 13g 14g 43g
## 20 1077393 Easter Sugar … http… jpg 573 8g 25g 80g
## # … with abbreviated variable names ¹imageType, ²calories
In this section we have created two new variables and added them to the
data frame and also created three contingency tables explaining the
association between the variables of the dataframes. The nutrients
tibble is extracted from the AmericanCuisineInfo data frame.
In the below code chunk the information extracted from the API does not
have information related to total price of the recipe and proteins
per serving. Some variables of interest provided in the data was
information about price per serving, total servings and
total_protein. We used these variables and created two new variables
totalprice and protein_per_serving.
It was interesting to find out that total price provided in the API was
in pennies. So, for total price we calculated this by multiplying
price per serving and servings then dividing by 100 to get a dollar
amount for total price. For protein per serving we divided amount of
protein in a recipe by total number of servings the recipe made.
Some nutrients that we were interested in were calories, protein, fat, and carbohydrates.
# Example for data exploration
# American cuisine info object that was created above
CuisineInfo <- AmericanCuisineInfo
# viewing nutrients data frame nested inside results list
View(CuisineInfo$results$nutrition$nutrients[[1]])
# want to look at calories, protein, carbs, fat
# bind_rows - binding row wise for all lists
# nutrients is a list of dataframes
nutrients <- bind_rows(CuisineInfo$results$nutrition$nutrients, .id="df") %>%
as_tibble() %>%
filter (name %in% c("Calories", "Protein", "Fat", "Carbohydrates")) %>%
pivot_wider(id_cols=df, names_from=c(name, unit), values_from=amount) %>%
bind_cols (CuisineInfo$results[,c("id","title", "vegetarian", "vegan", "pricePerServing", "readyInMinutes", "servings")]) %>%
mutate (totalprice = (pricePerServing*servings)/100, protein_per_serving = (Protein_g/servings ) )
# results is a data frame that we can use to do EDA
# cost per serving is in pennies.
# I converted total price/cost to make the recipe in USD
# output nutrients tibble
nutrients## # A tibble: 100 × 14
## df Calorie…¹ Fat_g Carbo…² Prote…³ id title veget…⁴ vegan
## <chr> <dbl> <dbl> <dbl> <dbl> <int> <chr> <lgl> <lgl>
## 1 1 590. 21.9 80.5 23.4 6.43e5 Fala… TRUE TRUE
## 2 2 393. 29.0 7.17 26.9 6.47e5 Salm… FALSE FALSE
## 3 3 455. 25.9 25.6 33.7 6.39e5 Chil… FALSE FALSE
## 4 4 479. 8.77 85.0 18.7 6.51e5 Masa… FALSE FALSE
## 5 5 450. 17.5 11.7 56.9 6.60e5 Slow… FALSE FALSE
## 6 6 475. 18.2 27.7 49.0 6.38e5 Chee… FALSE FALSE
## 7 7 971. 69.6 34.0 50.6 7.76e5 Gorg… FALSE FALSE
## 8 8 755. 20.5 58.9 82.6 6.57e5 Pull… FALSE FALSE
## 9 9 588. 23.9 28.6 65.4 6.62e5 Supe… FALSE FALSE
## 10 10 311. 8.83 33.8 23.2 1.10e6 Salm… FALSE FALSE
## # … with 90 more rows, 5 more variables: pricePerServing <dbl>,
## # readyInMinutes <int>, servings <int>, totalprice <dbl>,
## # protein_per_serving <dbl>, and abbreviated variable names
## # ¹Calories_kcal, ²Carbohydrates_g, ³Protein_g, ⁴vegetarian
In the below chunk of code, we created contingency tables to find the
associations between vegan and vegetarian, Source and Healthy,
Source and Expensive.
In the vegan and vegetarian contingency table, we see that there are
5 vegan recipes in the American Cuisine with fruit being the
ingredient that is being searched. So, we can conclude that some
vegetarian recipes are vegan as well. This is as expected.
In the Source and Healthy contingency table, we could see that most of
the unhealthy food recipes are from Foodista source. Now we know
which website to avoid for healthy food options.
In the Source and Expensive contingency table, we added one more
variable to the dataframe called expensive, which was a category we
defined if the price per serving was higher than Q3. Here, we can see
that the recipes suggested by Foodista has the maximum number of
expensive recipes, but also suggests maximum number of
inexpensive recipes.
# make 3 contingency tables
# vegetarian, vegan
VegConTbl <- base::table("Vegan" = nutrients$vegan, "Vegetarian" = nutrients$vegetarian)
VegConTbl## Vegetarian
## Vegan FALSE TRUE
## FALSE 59 36
## TRUE 0 5
# sourceName and Healthy
sourceHealthyConTbl <- base::table("Source" = CuisineInfo$results$sourceName, "Healthy" = CuisineInfo$results$veryHealthy )
sourceHealthyConTbl## Healthy
## Source FALSE TRUE
## A Southern Grace 1 0
## Afrolems 2 0
## blogspot.com 1 0
## Foodista 70 1
## foodista.com 20 1
## Pink When 2 0
## spoonacular 2 0
#create category
#price expensive
#price in pennies
new_CuisineQuery <- CuisineInfo$results %>%
mutate ( "expensive" = pricePerServing >= quantile(CuisineInfo$results$pricePerServing, 0.75))
sourceExpensivecontTable <- base::table("Source" = CuisineInfo$results$sourceName, "Expensive" = new_CuisineQuery$expensive )
sourceExpensivecontTable## Expensive
## Source FALSE TRUE
## A Southern Grace 1 0
## Afrolems 1 1
## blogspot.com 0 1
## Foodista 60 11
## foodista.com 11 10
## Pink When 1 1
## spoonacular 1 1
In this section we are summarizing the data by using the following
functions mean(), median(), desc(), select(), summarise(), and
sppaly().
In the following code chunk, for the mean_protein object, we did a
mean price breakdown with a group_by() function with vegetarian and
non vegetarian. In the highProteinRecipes object, we can see that
Cinnamon French Toast Sticks has higher protein than Turkey
Burgers with Slaw even though it is a vegetarian recipe. This is
surprising, but also makes sense due to the protein content in eggs.
# mean price breakdown by group - vegetarian and non vegetarian
# from data above in other code chunk
mean_protein <- nutrients %>%
group_by (vegetarian) %>%
summarise("Count" = n(), "Mean" = mean(Protein_g), "SD" = sd(Protein_g), "Med." = median(Protein_g), "Min."=min(Protein_g), "Max." = max(Protein_g))
# protein content per serving size (for the recipes)
# view the data to see what the protein per serving is.
highProteinRecipes <- nutrients %>%
select (vegetarian, vegan, title, protein_per_serving, Protein_g, servings) %>%
arrange (desc(protein_per_serving)) %>%
head(5) In the code chunk below we have created objects by fetching the
information of multiple cuisines from the Spoonacular API. paste() is
used along with collapse to pass as a parameter to the
my_multiCuisine_API() custom function. The code chunk returns a tibble
containing the number of recipes for each cuisine in the data that we
extracted for all the ingredients and given protein range. It can be
seen that the number of American cuisine recipes are much higher
in comparison to Chinese and Latin American cuisines.
# Vector of multiple cuisines
cuisinetxt <- c("American", "chinese", "Latin American")
# Using paste to create a text string of cuisine types to fetch information
cuisineComb <- paste(cuisinetxt, collapse = ",")
# Calling my_multiCuisine_API function to get the combined information
MultiCuisineInfoObj <- my_multiCuisine_API(cuisinetxt = cuisineComb, ingredient = "", nutrientinfo =
"TRUE", number=100, diet = "", minProtein = 10, maxProtein = 100)
# Object with parsed data collected from API
spoonacularTemp <- fromJSON(rawToChar(MultiCuisineInfoObj$content))
# Summarizing n counts based on cuisines
MultiCuisineInfo <- spoonacularTemp$results %>%
group_by(cuisines) %>%
summarize(count=n()) %>%
filter(cuisines != "character(0)")
# unlist the list of cuisine names and combining it into single name using apply
cuisineInfoTemp <- apply(MultiCuisineInfo, 1, function(y) sapply(y, function(x) paste(unlist(x),
collapse="_")))
# Taking transpose of the df to makes it useful for plotting
cuisineInfoTempTranspose <- t(cuisineInfoTemp)
# Converting it to tibble
MultiCuisineInfoTbl <- as_tibble(cuisineInfoTempTranspose)
# The count variable was character, used sapply to conver it into numeric
MultiCuisineInfoTbl$count <- sapply(MultiCuisineInfoTbl$count, as.numeric)
# Multicuisine output tibble
MultiCuisineInfoTbl## # A tibble: 3 × 2
## cuisines count
## <chr> <dbl>
## 1 Chinese_Asian 11
## 2 American 80
## 3 South American_Latin American 9
In this section, we did data exploration through multiple plots - like bar plot, histogram, boxplot, scatter plot, and scatter plot with linear model.
In the code chunk below, we have used a bar plot to show the recipe
count variations among the different cuisines. In this code chunk we
have used geom_bar() for bar plot and theme() to tilt the x-axis
names as they are overlapping. It can be seen from the plot that
American cuisine has highest number of recipes with all the
ingredients with protein range between 10-100.
barPlot <- function(df, xVar, yVar){
g <- ggplot(data = df , aes(x = xVar, y = yVar))
g + geom_col(aes(fill = as.factor(xVar))) + scale_fill_discrete(name = "Type of Cuisine") +
xlab("Cuisines") + ylab("Count") +
ggtitle("Barplot For Cuisines Type Counts") + theme(axis.text.x = element_text(angle = 18))
}
barPlot(MultiCuisineInfoTbl,MultiCuisineInfoTbl$cuisines,MultiCuisineInfoTbl$count)
In this histogram, we have plotted for total price variation for
American cuisine. We have used geom_histogram() for the histogram
plot and added another layer to it using geom_vline(), showing the
mean of total prices on the histogram plot. The histogram plots does
not look like a normal distribution. It looks like it skewed towards the
right. Maybe in future data exploration, we can transform the data into
something more familiar and be able to draw conclusions from that
familiar distribution.
histPlot <- function(df, xVar, bin_width = 2){
g <- ggplot(data = df , aes(x = xVar))
g + geom_histogram(binwidth = bin_width, color="blue", fill="white") +
geom_vline(aes(xintercept=mean(xVar)),
color="red", linetype="dashed", size=1) +
xlab("total prices") + ggtitle("Histogram for totalprices")
}
histPlot(nutrients, nutrients$totalprice)
In this box plot we
are looking to determine how much protein is present in the vegetarian
and non vegetarian meals. We have used geom_boxplot() for this scatter
plot and added a jitter layer using geom_point(). When looking at the
protein content in the vegetarian and non vegetarian recipes, protein is
not non-existent in vegetarian meals. This is a common misconception.
The number of recipes for each category also do not differ much. Also,
we see that the protein content in non vegetarian meals in generally
much greater than the protein content in vegetarian meals. Even the
outliers are only a little more than the median protein content in the
non vegetarian meals. Something to consider is how many serving size
each recipe is. This may contribute to a large difference in protein.
# boxplot()
# setting the stage to add layers
# how much protein is in vegetarian meals
g <- ggplot(nutrients, aes(x = vegetarian, y = Protein_g))
# boxplot for Protein by Vegetarian
g + geom_boxplot() +
labs(title = paste0("Boxplot for Protein (g) vs. Vegetarian Recipes (", CuisineInfo$cuisinetxt, " Cuisine)"), x = "Vegetarian Recipe", y = "Protein (g)") +
geom_point(aes(color=vegetarian), position="jitter") 
In this box plot we
will be looking at how much protein is in vegetarian meals per serving.
We have used geom_boxplot() for this scatter plot and added a jitter
layer using geom_point(). From the plot, interestingly enough, when
broken down by serving, the protein content in vegetarian meals (for
some) have a higher protein content then the non vegetarian meals.
Compared to the boxplot above this is an interesting finding.
# boxplot()
# setting the stage to add layers
# how much protein is in vegetarian meals per serving to see if there is any difference...
g <- ggplot(nutrients, aes(x = vegetarian, y = protein_per_serving))
# boxplot for Protein by Vegetarian
g + geom_boxplot() +
labs(title = paste0("Boxplot for Protein (g) per Serving vs. Vegetarian Recipes (", CuisineInfo$cuisinetxt, " Cuisine)"), x = "Vegetarian Recipe", y = "Protein (g) / Serving") +
geom_point(aes(color=vegetarian), position="jitter")
In this scatter
plot, we tried to see the relationship between protein and carbohydrates
of a particular cuisine. Here, geom_point() was used and we can see
that there is no real pattern to be observed between carbohydrates
and protein. A few points may indicate increase when either one is
increased, but over all protein stays low no matter how the carbohydrate
content varies.
# seperate scatter plot
# Carbohydrates_g vs. Protein_g
# setting the stage to add layers
g <- ggplot(nutrients, aes(x = Carbohydrates_g, y = Protein_g))
# scatter plot with linear regression line for carbs by protein
g + geom_point(aes(color=Protein_g), position="jitter") +
labs(title = "Scatterplot for Carbohydrates (g) vs. Protein (g)", x="Carbohydrates (g)", y="Protein (g)") 
In this scatter
plot, we have added a layer for the regression line using the
geom_smooth() function. After seeing the plot, we can see that the
regression line basically has No change in protein (g) at any amount
of carbohydrates. This is similar to the plot above.
#setting the stage to add layers
g <- ggplot(nutrients, aes(x = Carbohydrates_g, y = Protein_g))
#scatter plot with linear regression line for carbs by protein
g + geom_point(aes(color=Protein_g), position="jitter") +
labs(title = "Carbohydrates (g) vs. Protein (g)", x="Carbohydrates (g)", y="Protein (g)") +
geom_smooth(method = lm) 
# Wrap up
To summarize everything that we did in this vignette please see the following:
-
Built custom functions to interact with some of the Spoonacular Food and Recipe API endpoints like ComplexSearch and findByNutrients
-
Retrieved some of the data based on different user modifications
-
Did exploratory data analysis (EDA):
-
Using contingency tables
-
Numerical summaries like mean, median, min, max, standard standard deviaiton
-
Data visualization methods like plotting – bar plot, box plot, histogram, scatter plot, scatter plots with linear models overlayed.
-
Some interesting findings for our specific vingette examples:
-
Some vegetarian recipes are vegan as well
-
Most of the unhealthy food recipes are from the Foodista source
-
Recipes suggested by Foodista has the maximum number of expensive recipes, but also suggests maximum number of inexpensive recipes
-
Cinnamon French Toast Sticks has a higher protein content than a non vegetarian meals like the Turkey Burgers with Slaw, even though it is a vegetarian recipe. This is surprising, but also makes sense due to the high protein content in eggs
-
It can be seen that the number of American cuisine recipes are much higher in comparison to Chinese and Latin American cuisines for our vignette for the recipes containing fruit.
-
The protein content in non vegetarian meals in generally much greater than the protein content in vegetarian meals as expected
-
When broken down by serving, the protein content in vegetarian meals (for some) have a higher protein content then the non vegetarian meals
-
There is no real pattern to be observed between carbohydrates and protein
-
The regression line basically has No change in protein (g) at any amount of carbohydrates, confirming out plot that is only a scattet plot between protein and carbohydrates
We hope this vignette will help the user to interact with spoonacular API and do the required exploratory analysis.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent