Order Portal-Flow

Implementation Architecture

Workflow

• Order can be ingested from the mobile order dashboard using API ingestion or from pos order in the Kafka event for the pipeline.

• On top of this data, preparation is done for data transformation. Further data enrichment is done.

• With DS lab, customer segmentation, customer lifetime value & forecasting model deployed in the pipeline. The churn prediction model is used as API from DS Lab.

• The first flow on top shows how an order is getting from the mobile order to the kitchen dashboard with a kitchen push.

• The kitchen dashboard shows churn prediction using a model-like API. Once the order is processed from the kitchen dashboard it’s processed further to the delivery dashboard.

• In the second flow, incoming orders get split based on the order type. It processes further based on online order type. Then this recommendation model from DS Lab will push the recommendation offer to mobile order dashboard-based user past purchase & food order similarity etc.

• The third flow is about customer campaigns. Based on customer segmentation through the pipeline, it sends a coupon to customers via mail.

• So, all the data points during this entire process are collected & stored in a different collection of MongoDB databases.

Repository of DS Lab Models

• Customer segmentation + Customer similarity: Algorithm that segments customers according to customer buying behaviour (RFM values), as well as using a similarity score to return the top 5 most similar customers for each customer.

• Recommendation System: An algorithm that produces recommended products for input customers based on their similarity to other customers.

• Sales Forecasting: Use a Data Science model to forecast sales on a per-week basis based on the given data of sales over time.

• Churn Prediction: Use various Customer data to predict whether they churn or not on a per-quarter basis.

• Retail Recommend MBA: Contains Market Basket Analysis (MBA) of the product data.

• MBA Sales Data with Visualization: Contains various visualizations for the MBA of sold products.

• Product Analysis: Contains location and franchise-based analytics of product sale