Supply Chain Twin Demo - Companion Deck
0 likes•176 views
The document is a presentation deck on building a supply chain twin using Neo4j and Google technologies. It discusses how supply chain data can be modeled as a graph and stored in Neo4j to power use cases like identifying product and part shortfalls, evaluating supply chain risk, and enabling scenario planning. The deck outlines an architecture that ingests supply chain data from Google BigQuery into Neo4j, then leverages Neo4j technologies like Graph Data Science, Bloom, and Keymaker to operationalize queries and deliver insights to applications.
Supply Chain Twin Demo - Companion Deck
- 1. Neo4j, Inc. All rights reserved 2022 Neo4j, Inc. All rights reserved 2022 1 Supply Chain Twin Demo Companion Deck Nicolas Rouyer Sales Engineer, Neo4j
- 2. Neo4j, Inc. All rights reserved 2022 2 Graphs in Supply Chain Management & BOM
- 3. Neo4j, Inc. All rights reserved 2022 3 Worldwide supply chain problems
- 4. Neo4j, Inc. All rights reserved 2022 Organizations Facing Challenges Across Supply Chains Worsened by the Pandemic Planning Sourcing Production Sales Warehouse & Distribution 69% 74% 68% 69% 67% Difficulty planning due to lack of information on impacted suppliers Shortages of critical parts / materials Difficulties in balancing stock between warehouses Difficulties reconfiguring product lines Lost sales due to stock outs Percentage of organizations that faced significant challenges in each area Source: Capgemini Research Institute, Supply Chain Survey, August-September 2020, N=1,000 organizations
- 5. Neo4j, Inc. All rights reserved 2022 5 Worldwide Supply Chain Problems are Data Problems... 75% 938 of Fortune 1000 had a T1-2 supplier impacted by the pandemic Good forecasts are just less bad forecasts of executives aren’t confident in their data quality
- 6. Neo4j, Inc. All rights reserved 2022 6 Supply Chain is a Graph
- 7. Neo4j, Inc. All rights reserved 2022 7 Supply Chain is a Graph ● Customers ● Orders ● Employee ● Suppliers ● Materials ● Products ● Plants ● Distribution Centers ● Shipments ● Global News ● Etc…
- 8. Neo4j, Inc. All rights reserved 2022 8 Supply Chain is a Graph Because of its connected nature, supply chain is a massive data challenge
- 9. Neo4j, Inc. All rights reserved 2022 9 Demand Products Manufacturing Suppliers Order Fulfillment Transportation Weather Geospatial Third Party Leveraging Cross-Silo Connections The property Graph Model
- 10. Neo4j, Inc. All rights reserved 2022 Google + Neo4j Supply Chain Twin Architecture BigQuery Supply Chain Twin L2 Data Dataflow Graph Database Bloom Graph Data Science Keymaker Framework Supply Chain Twin Application Analysts Supply Chain Service Source data is present in Google supply chain twin L2 canonical and GDELT public datastore. Google Dataflow is used for one time bulk and batch data processing for consuming Google supply chain twin L2 canonical data by executing Google BigQuery and loading that data into Neo4j Database after transformation using Neo4j Spark Connector integration. Neo4j database enables you to store, query, analyze, and manage highly connected data. Neo4j supports Apache Spark, Kafka Streams, BI and database drivers integration to load/read data into/from Neo4j database. Neo4j Graph Data Science (GDS) is a connected data analytics and machine learning platform that helps you understand the connections in big data to answer critical questions and improve predictions. Neo4j Bloom data visualization tool allows analytics team to freely explore graph data to validate findings and improve explainability. Here web application also embeds its UI for an easier accessibility of contextual data. 1 2 3 Neo4j Keymaker is a data model agnostic tool designed to help organizations operationalize their graph based analytical queries. Here it executes product risk score, product parts shortfall and other queries to serve data insights to web application through its GraphQL endpoints. 4 Application Service is implemented using Neo4j GraphQL Library to access Neo4j database data using low code approach. Supply Chain Application allows customer to find additional insights of their supply chain data which gets enriched with highly connected data insights using Neo4j Graph database and Data Science services. GDELT Data 1 2 3 4 5 6 5 6 Google Applications Google Applications can integrate to consume this connected data insights by invoking Keymaker or Supply Chain Application service GraphQL endpoints. 7 7 Products Solution Accelerators Flex Templates Solutions Workbench A Iterative Graph Modeling. (Design time). A
- 11. Neo4j, Inc. All rights reserved 2022 11 Google Supply Chain Twin L2 Data Model
- 12. Neo4j, Inc. All rights reserved 2022 12 Neo4j Supply Chain Twin - Graph Data Model
- 13. Neo4j, Inc. All rights reserved 2022 13 Orders, Fulfillment and Product Shortfall • Total orders • Orders unable to be fulfilled • Products and Parts shortfall against the orders
- 14. Neo4j, Inc. All rights reserved 2022 14 Best Selling Products, Common Parts • Current Best Sellers • Common Parts used by best sellers • Low inventory common parts • Which of these common parts can be made more of based on inventory levels multiple levels deep • Legal Entities for low inventory parts and their location
- 15. Neo4j, Inc. All rights reserved 2022 15 Part Shortages per Order (multi-level + tree) • Part Shortages per order for each level of the part • Multi-level ◦ Shows how many more parts can be made using low level parts • Tree ◦ Shows suppliers from top-level products ordered by shortfall
- 16. Neo4j, Inc. All rights reserved 2022 16 Part Shortages per Order • Graph traversals make it easy to traverse the supply chain - looking for parts made of other parts • Inventory information can be looked up as part of the traversal to calculate shortfalls Part graph traversals Parts Inventory
- 17. Neo4j, Inc. All rights reserved 2022 17 Part Shortages per Order (Sankey) • Sankey diagram looks where the part is used further up the supply chain
- 18. Neo4j, Inc. All rights reserved 2022 18 Risk Scoring based on Orders, Inventory, Part Centrality and Location
- 19. Neo4j, Inc. All rights reserved 2022 19 Evaluate Supply Chain Risk • Analyze the most critical parts and calculate the part centrality risk score using betweenness centrality • Calculate location risk score based on the coordinates of the manufacturer’s address • Calculate order risk score based on current inventory levels and order quantities Provide actionable insights by simulating more inventory and order and modifying risk weightage on the fly. High Centrality Product News Events Order Fulfillment Risk
- 20. Neo4j, Inc. All rights reserved 2022 20 Additional Use Cases Demand Sensing • Forecasting across Products, seasons/time periods, Location, etc. Channel Shaping • Analyzing orders by customer / partner, frequency, trends, etc. Supply Management • Inventory forecasts • Demand based Inventory replenishment • Alternate Supplier Dynamic Risk Alerting • News monitoring and event based risk calculations and alerting • Multi factor risk calculations Business Planning • Scenario based planning and analysis • Balance demand, supply, inventory and financial targets