Each cuisine is distinct because of a small set of distinct ingredients. We can't surface these by looking at the most popular ingredients, since they're the bread-and-butter ingredients of cooking like salt or sugar.

Another metric that can aid us in this quest is the term frequency–inverse document frequency (TFIDF), a numerical statistic that is intended to reflect how important a word (ingredient) is to a document (cuisine) in a collection or corpus (recipes).

We already precomputed the tf_idf for you and created an enriched dataset named recipes_enriched. Your goal is to create a Shiny app that displays a horizontal bar plot of the top distinctive ingredients in a cuisine, as measured by tf_idf.

An app displaying an interactive horizontal bar plot of top ingredients by chosen cuisine

You will use a reactive expression to encapsulate the computations and let the plotting code focus only on creating the plot. This is good programming practice and helps create modular and performant Shiny apps.

We have loaded the packages shiny, dplyr, ggplot2 and plotly. Here are two handy snippets to filter for the top recipes by cuisine and create a horizontal bar plot. You can modify it appropriately.

top_ingredients <- recipes_enriched %>% 
  filter(cuisine == 'greek') %>% 
  arrange(desc(tf_idf)) %>% 
  head(5) 

ggplot(top_ingredients, aes(x = ingredient, y = tf_idf)) +
  geom_col() +
  coord_flip()

Wrap up video

Create Your Free Account