What is Datatron? from Harish Doddi on Vimeo.
When it come to Data Science, Datatron provides an excellent Model Life Cycle Management Platform.
Facilitating product ionizing and monitoring and Machine Learning (and Deep Learning) models. Datatron’s hyper-local Machine Learning and deep learning models create a more intimate relationship between consumers and brands.
The models and the team reveal the consumers who will demonstrably spend in proposed store locations. Datatron builds a geo-hashed expected revenue model (and map). This map is interactive and incorporates tens of millions of geo-hashes, each with unique revenue estimates. When an analyst clicks on a geo-hash, he/she gets information related to “What if they open a store in that location specified by the geo-hash?”
The output comprises of things like “Expected revenue of the store which is associated with a default set of variables (deep learning input variables) tied to a Probability vector (i.e. how confident the model is to get this expected revenue with these variables). A Deep Learning model requires and can be enhanced tapping many data sources 1. First Party Data. This is defined as the chain’s own data. a. Shopper data (Required): This data typically includes user id, transactions done by the user, type of good bought, storied, etc. b. Products data (Optional): Any information relating to products getting’s sold c. CRM data (Optional): Any information users share with the chain. d. Subscription data (Optional): E.g. Newsletters, loyalty programs. e. Social data (Optional): Data pulled on people who interact with the chain’s brand. f. Cross-platform data (Optional): Data pulled from mobile or web apps. g. Offers behavior data (Optional): This data typically includes store discounts, offers, sales, etc.
Please contact with you thoughts and questions Theo Olison 703-200-8534 [email protected]