HACKER Q&A
📣 LegendaApps

Convert Product Descriptions into Metadata?


Hi,

I'm looking for a tool that can extract metadata and categorize products based on descriptions. So if I'm selling 10 different fragrances for example, this program should analyze the text and identify that the [colors] are different.

I think a YC startup does something like this. Anyone have pointers to code or what this domain is called?

Thanks!

  👤 Jugurtha Accepted Answer ✓
Are you referring to ontology[1] and Semantic Web[2]? Did you take a look at https://schema.org/, particularly https://schema.org/Product. Practically all ecommerce websites use it.

[1]: https://en.wikipedia.org/wiki/Ontology_(information_science)

[2]: https://en.wikipedia.org/wiki/Semantic_Web

👤 billconan
I'm building a solution for our company now.

In the end, it's a mix of human labor and machine learning. We use keyword matching to categorize the product and then NLP to understand some of the fields, such as product dimensions.

This domain is called data ETL. Which yc company does it? I want to know it too.

👤 2rsf
You need to give us some more context, what do you mean by "analyze" ? how will the "program" get the data ? Is it textual analysis ?