One of our team members works with an organization called Health Leads. Health Leads is a student volunteer based organization that helps low income populations find community resources that can improve health outcomes.
What it does
Low income populations tend to be less educated when it comes to nutrition. Advertisements are misleading and precious income is commonly spent inefficiently on food that provides neither adequate macronutrients or micronutrients to the consumer. Our app takes user inputs regarding income and food preferences and generates a grocery list suggesting foods for healthy eating. These suggestions could be talked over with a physician or Health Leads advocate and modified accordingly.
How we built it
- We constructed a database of common foods with their associated unit price using FDA API and Python to automatically extract the data.
- We used simplex method in R to optimize the list of foods given the constraints of the user (food preferences, income, family size). This outputs a list of food items (and their quantities) that would satisfy a family's daily nutrition needs
- We used Shiny (a web application framework in R) to build a prototype of our application.
Challenges we ran into
- Normalizing food/cost data
- Learning basic syntax of Shiny and R in the limited time frame of a hackathon
What's next for What's for Dinner?
- Upgrade the optimization algorithm to be more sensitive to micronutrient profiles of foods
- Utilize Walmart API (get price per pound) to link Food ID from USDA to create a giant self-updating database
- Improve the UI to be more intuitive and visually appealing
- Create an interface with USDA’s Interactive DRI for Healthcare Professionals