6 weeks project (including Fletcher):
Goal : Beauty product recommendation.
1) Get Amazon Data and Nordstrom Data
2) Create a better sorting algorithm
(based on number of stars : Laplace transformation, and weight 1-start review heavier)
3) Natural Language processing : TF-IDF, Document Summarization, Sentiment Analysis Graph (vs Star-Graph)
* Build product-user-profile : skin type and sensitivity
4) Build an application where the user can type in specify requirement for the product.
Based on text comparison, return a maximum of 3 products as recommendations.
UP AND RUNNING : Beauty3.ninja (currently under maintenance)
Website name explained:
beauty3 : We only recommend 3 beauty products at a times (yay, no overwhelming amounts of products)
.ninja : Because just as ninja, we are highly skilled and precise. (secretly, as always come with "ninja", is because it's awesome :D)