Retail Supply Chain

**Overview**

CPG is an exciting space for analytics, but analysts, data engineers and data scientists at CPGs are often lacking easy-to-access and up-to-date data. The data needed does exist, but it's often locked away in clunky retailer reporting and file sharing systems that are hard to integrate with modern data stacks.

Crisp connects to over 40 leading U.S. retailers and distributors. Our self service data pipeline ingests a live-feed of data from retail vendor portals to bring you real-time data directly into Databricks that is normalized, reliable, and ready to use — so you never have to download a report again. Of course, Crisp also integrates modern retailer and distributor systems such as Walmart Luminate and UNFI Insights.

**Types of supply chain data provided via Crisp:**

- **DC fill rates.** Delivery performance from DCs to stores and inbound to DCs.
- **DC Inventory.** DC on hand and on order inventory.
- **DC to store shipments.** Shipments from distributor or retailer DCs to stores
- **Store Inventory.** On Hand inventory and in-stock levels.
- **Much more.** Crisp also provides many other datasets, depending on source availability. For example, age of inventory and spoilage risk, chargebacks and deductions.

Not only does Crisp automate the ingestion of retailer and distributor data, we also can take care of cleaning and organizing the data so you can spend less time on data prep and more time on actual analysis.

**The Crisp Platform provides three different types of data, with different levels of pre-processing:**

**Source data.** For when you are developing a replacement for existing data pipelines and / or want to do your own data modeling. The source data schemata closely resembles what the upstream data source provides, so you can leverage any pre-existing modeling experience with this data.
**Normalized data.** For when you are developing retailer or distributor specific solutions and want to leverage source-specific attributes. The data model is a fully normalized schema and naming conventions are consistent across all data sources, making it easy to mix and match retailers in analyses.
**Harmonized data.** For when you are developing cross-retailer solutions, and it’s acceptable to use a narrower subset of data that is common across multiple data sources. The data model is built on the normalized model and combines common data elements, such as POS and inventory into a harmonized model across multiple retailers and distributors.

With Crisp, you can easily mix and match data from the models as you see fit.

**The applications of the data are vast, but some common use cases are:**

- **Reduce out of stock.** Use retailer inventory and sales data to easily detect out of stock and customer availability issues.
- **Increase distribution.** Use shipment data from distributors or retailers to identify Gaps in distribution.
- **Optimize DC Inventory.** Use DC order and inventory data to optimize DC replenishment and maximize fill rates.
- **Improve demand forecasting.** Combine retailer inventory and sales to generate demand forecasts and inventory allocation plans.

搜集汇总

背景与挑战

背景概述

该数据集通过Crisp平台整合美国40余家零售商的实时供应链数据，提供原始数据、标准化数据和跨零售商统一数据三种层级，支持库存管理、分销优化和需求预测等供应链分析场景。数据经过清洗处理可直接用于分析。