SALES performance prediction with aws EMR

State of business prior to engagement
Mabaya helps advertisers and brands to present products by putting them in front of relevant shoppers at the moment of truth. To carefully predict which users are likely to purchase a certain brand or product, Mabaya needs to precisely analyze large amounts of shopping information in short time.

Problem statement
Reliably collecting, processing, transforming and analyzing large amounts of merchant provided information could be quite challenging task, given the amount of the data. Previous solutions based on off-the-shelf Apache Hadoop felt short because the solution wasn’t cost-efficient enough.

Technical solution
By utilizing AWS Elastic MapReduce, Mabaya can now effortlessly process millions of user's profiles every day and provide precisely targeted recommendations and runs machine learning algorithms to build perfectly accurate sales.

Data Analysts are using Apache Zeppelin along with AWS MapReduce to work on new models as well as analyze performance of current models and potentially optimize them.


For advertisers and brands, we help you present your products by putting them in front of relevant shoppers at the moment of truth  – when they are in the store, ready to buy. It is every merchandiser’s dream.

It’s simple: We obtain premium placements for your products in the online store, where potential shoppers can actually see them. Greater visibility means increased engagement, and increased engagement means more sales.