System overview

  A system that predicts sales of the entire store and each product by utilizing POS data of the retail business, weather information acquired from the outside, and trend information. It can also be used for past sales analysis and trend analysis.

 

Client

  Stationery sales

 

Effect after system introduction

   We were able to place orders at the right time, reduce defective inventory and missed sales, and improve profits by about 5%.

 

Technology

• Prediction engine: RNN model, LightGBM model

• Framework: Keras, Flask

• Language: Python

​

Development scale

  2 people x 4 months