Data mining in the cloud

Do you need structured data? Our Artificial intelligence will gather it from the web.

Jill runs a shopping search engine. Her users want to search products by params. e.g. sort digital cameras by resolution:

Digital camera Type Res.
Canon PS S100 Compact ?
Nikon S3200 ? 24 Mpx
Canon EOS M DSLR 18 Mpx
Samsung NX210 DSLR ?
Nikon Coolpix S800C ? 16 Mpx

She has some data but collecting it is:

  • expensive
  • labor intensive
  • error-prone

 

Digital camera Type Res.
Canon PS S100 Compact 12 Mpx
Nikon S3200 DSLR 24 Mpx
Canon EOS M DSLR 18 Mpx
Samsung NX210 DSLR 20 Mpx
Nikon Coolpix S800C Compact 16 Mpx

She tries MagicTable

  • we automatically learn structure of her data
  • crawl the relevant information from the web
  • fill in the blanks

Structure analysis

Our algorithm examines your data

  • strings/categories - What values are possible?
  • numbers and units - What range is usual?

Focused crawling
& natural language

We gather relevant web pages by our crawler.
We analyse the texts and identity information.

Machine learning

We combine the data from all the web pages and from your own data to predict the missing values.