TARGETED MESSAGING ABOUT FOOD STORAGE IN SOCIAL MEDIA POSTS

With current global food waste rates being at an extreme, the question arises of how do restaurateurs, particularly ones searching for the perfect Instagram picture to post, contribute to global food waste? The problem is more relevant now, seeing that users are so focused on trying the trendiest foods, however not realizing the amount of food they waste. The solution to this is to educate the users about ways that they may be able to store their leftovers to be consumed at a later time. This research project aims at creating a machine learning program using Python and Microsoft Computer Vision in order to process Instagram pictures and Twitter tweets. The machine learning program will determine the food presented in the post and display an informational message on how that food can be stored for later consumption. A dataset, provided and published by the U.S. Department of Agriculture, contains a variety of foods and ways those can be stored through three means: pantry, refrigeration, and freezing. The method of this research project included reading twitter tweets as well as processing Instagram images to determine what food was being displayed or talked about. The program then analyzed the dataset and recognized the food item. If the food was present in the dataset, the program displayed leftover storage options. As a result, the program accurately found and identified the food from the image. My long-term career objective is to focus on projects that would bring some sort of benefit to humanity. The impact of our findings informs users on Instagram and Twitter on ways to preserve leftovers. This ultimately reduces food waste. Our hope is that these results cause our social media users to consider how their actions might impact society negatively and how we can use technology to solve societal problems.