TechLabs Digital Shaper Program
I got accepted as a top 50 students in Denmark into a 4-month data science bootcamp at TechLabs, Copenhagen, Denmark.
My Contributions
As a data analyst, I used python to come up with machine learning model to classify food into vegan/non-vegan and vegetarian/non-vegetarian from their ingredients. In this showcase, I will demonstrate my methodology for only vegan (since the vegetarian also has the same process). You can see my code at this Google Colaboratory link.
https://colab.research.google.com/drive/1A1YG5LfBH9hv08HWwHKLzMDXGg5oS0XC?usp=sharing
In addition to my role as a data analyst, as a product manager, I oversaw weekly meetings with three engineers from Russia, Iceland, and Denmark. The Double Diamond approach was what I used as a UX/UI designer to identify, define, create, and deliver the product.
Data Analytic Processes
-
Data Requirement Gathering Since the ingredients of the food are the indicators if the food is vegan or not, we want some test dataset that has the food name, their ingredient name, ingredient ID, and tag (target value).
-
Data Collection We found this open source data sets in Kaggle, thanks to Bodhisattwa Prasad Majumder, Shuyang Li, Jianmo Ni, and Julian McAuley.
https://www.kaggle.com/datasets/shuyangli94/food-com-recipes-and-user-interactions
-
Data Cleaning
-
Data Analysis Final Result: 0.9568333333333333
-
Conclusion We’re able to build a machine learning model that classify food from its ingredients (up to 20 ingredients) into vegan or non-vegan with 95.68% accuracy.
-
Reflection This task is a real challenge for me to work in a diverse group of new people, new place, and new disciplinary that I haven’t worked on before: data science, product management, and UX/UI at the same time. More about this project
Report:
UX Process:
https://www.notion.so/TechLabs-UX-Process-12cff3b645ed4c2e9806a02966f7bfc9?pvs=4
TechLabs Digital Shaper: