College students face many pain points cooking under time and budgetary constraints, further enhanced by the long study hours which leave many students hungry in the wee hours of the morning when shops are not open. Therefore we hope to alleviate some pain with our idea!

What it does

User inputs what ingredients are left in his fridge and program generates a few options of foods that can be cooked with these ingredients quickly, approximately in 15 minutes. Program further guides the user, presumably a time-pressed college student who is amateurish with cooking, through the cooking with written instruction, pictures and video tutorials. Bonus features include listing of recipes which user has some of but not all of the ingredients, so that college students have the flexibility to supplement fixed recipes with impromptu substitutes.

How I built it

We get our data from various popular recipe sites. Thereafter we curate it specifically for college students, specifically in the aspects of ease of preparation, simple ingredients and short cooking time.

Challenges I ran into

Integrating our back-end technologies with the front end and developing a large enough database. We also built a Machine Learning solution but could not integrate it with the front end for users.

Accomplishments that I'm proud of

We managed to make pleasing aesthetics for our user interface and packed an impactful demo.

What I learned

We picked up new technologies including TensorFlow, Neo4J and python backends. We also gained a newfound appreciation of front end designers.

What's next for College Cook

College Cook will leverage on the massive new datasets that have been made public in the recent months to develop a larger graph database comprising of more food and nutritional information for easy access by college students. Incorporate suitable technologies like machine learning to develop personalized recommendations. Incorporate aspects of the sharing economy for nearby classmates to cook together.


Ze Xuan Ong: Built the front end, integrated the database with the backend and front end. Taught team to use Neo4J. Urvi Agrawal: Compiled all the data and developed the graph database in Neo4J. Achieved image recognition for pictures of food in over 10 different classes using TensorFlow models of machine learning. Serano Tannason Ng: Built the back end using Python and front end using Javascript and Bootstrap. Achieved the integration of the machine learning model with the back-end. Brandon Pek: Developed pitch and features of practicality, brainstormed entrepreneurial expansions and possibility of financial scalability by leveraging on sharing economy. Incorporated humorous demonstrations of project concepts and features content through the video demo.

Share this project: