This document contains list of things that can be implemented on available products and user data.
Using groups:
We already have groups and subgroups of our products. There’re many things we can do with these groupings.
Finding outliers in any particular groups or subgroups. For example, we can answer questions such as, 1) Out of all yogurts, which one is the odd/special one from the others? What is the most unique yogurt in the group? 2) What makes it so special? Any huge differences in nutritional values/ingredients? 3) How much different is it from the most typical yogurt (absolute value/percentage difference)?
What product/food is the most typical/common according to nutrition/ingredients in any particular group?
Summarization of
…show more content…
Finding similarity between groups based on nutrition or ingredients. Such as, Cabbage and Kale are more similar than Cabbage and Apple. This can be used to build recommendation of similar products based on user preferences.
Stats of each products in comparison to other products in same group. Select any product in group and compare it to every other products to determine what makes that product different from all the other options. This will give important highlights about each product and can be used in an app.
Difference between any two products in groups. This difference can be calculated based on nutrition or
…show more content…
So, If user A likes Cokie XYZ and Cookie ABC is also similar to cookie XYZ, we can also recommend cookie ABC to user A. This similarity can be based on nutrition, ingredients, diet or any combination of these things.
Recommendation engine based on user data and product data with the use of collaborative filtering : If User A and User B is similar based on their preferences or search/browsing history and if User A likes Cookie XYZ then it’s very likely that User B will also like Cookie XYZ. So, we can recommend Cookie XYZ to User B. This can give use very important insights about underlying preferences of users. We can recommend better products to users and there’s an opportunity to tie up with big food brands If we have record of these preferences. Because by using this data, these brands can focus on targeted advertising and sales to decrease cost per customer acquisition.
Graph search:
Graph search on products. Such as, “All dairy free yogurts with protein > 10 grams” Or “ Keto food which is a snack and GMO-free with protein > 10 grams and fat < 4 grams”.
This search can be done on any combination of different criterias.