Amazon is an amazing company. They are an important part to many people’s lives. Amazon has quickly taken the place of brick and mortars stores everywhere. The one thing Amazon does not have verses bricks and mortar is the ability to shop with friends. There is nothing like using inspiration from a someone else to inspire your next purchase. Connecting people with like minded individuals is the next big step of virtual shopping experience. Amazon already has user profiles that shows previous reviews in one place. They have also seen the benefit of influencers using affiliate links to drive traffic to their site. This recommendation system will recommend users instead of products to give Amazon more of a social media feel so you can buy directly from the site, eliminating the need for third party sites.
Around 3 million of 10million entries from Clothe, Shoes, and Jewelry where use to create this recommendation system. Because of the size of the operating system all data was not able to be use. Random selection was chosen to select entries used.
- Surprised was used to test different recommendation systems.
- Baseline worked best on our sample size and used it to make the recommendations
- Implement with in Amazon web page as an additional footer
- Try out additional recommendation system on larger data set
- Reword front end of public profiles to make photos more prominent like on an Instagram page
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent