Product ranking without user interfering
On amazon.com or another popular e-commerce webpage for product, visitor can notice reviews and general opinion about product. Expected behavior of users after purchase and certain period of usage is to rank purchased product. This gives other visitors personal insight about product in which they are interested. Thus, product ranking is not hard to achieve when there is large number of users and large number of products. Challenge is to get product ranking for each product when there is small number of users and large number of products on website.
- Log in to post comments
2 answers
Product ranking using data from social networks
Product pages on e-commerce platforms usually have buttons which allows users to share products on social networks. Supported networks are usually most popular ones like Facebook, Twitter, Pinterest, Google+ etc.
Since each product has unique URL, we can use those URL and make HTTP calls/requests to different social networks APIs. Response from APIs consist count of how much time given URL is shared on that social network.
After we get response for several networks, we can calculate rough ,,social importance" by formula:
Social importance of product X = sum (importance of network Y * social count for product X on network Y) for Y < n
where n is total number of social networks we are inspecting.
After calculating social importance we are normalizing result to value from 0 to 100% and that gives us product ranking.
This calculation is then passed through certain ,,pipeline" which is performing each of those steps for each product. This can be done using Map/Reduce algorithm.
Further on as number of users is growing on website, and product ranking is changing, company can switch from our result to result based on user's reviews of product.
Taggings:
- Log in to post comments
Product ranking using data from social networks
Product pages on e-commerce platforms usually have buttons which allows users to share products on social networks. Supported networks are usually most popular ones like Facebook, Twitter, Pinterest, Google+ etc.
Since each product has unique URL, we can use those URL and make HTTP calls/requests to different social networks APIs. Response from APIs consist count of how much time given URL is shared on that social network.
After we get response for several networks, we can calculate rough ,,social importance" by formula:
Social importance of product X = sum (importance of network Y * social count for product X on network Y) for Y < n
where n is total number of social networks we are inspecting.
After calculating social importance we are normalizing result to value from 0 to 100% and that gives us product ranking.
This calculation is then passed through certain ,,pipeline" which is performing each of those steps for each product. This can be done using Map/Reduce algorithm.
Further on as number of users is growing on website, and product ranking is changing, company can switch from our result to result based on user's reviews of product.
Taggings:
- Log in to post comments