Time series analysis for shipping of outdoor equipment with Data Warehousing
Our client, a large seller, and distributor of outdoor equipment required a central platform for all data as a basis for Advanced Analytics of customer and business data. Based on the Google BigQuery Data Warehouse, Niologic implemented a time series analysis for shipping processes. Additionally, Niologic facilitated a machine learning algorithm for predicting supply shortages of service providers for Google Cloud AutoML and a Google Data Studio Dashboard for the monitoring of controlling on a daily basis.
Our client was using multiple platforms and data sources, such as e-commerce shop systems, SAP ERP, online APIs from marketing as well as internal systems like telephony. To improve the analytics potentials for their data and to lay a foundation for machine learning models, our customer wanted to integrate all data into a singular platform. After the integration into a data warehouse, Niologic was able to use this data for a combined view on customer interactions introducing machine learning models.
In addition, Niologic gathered business and shipping data from the shipping provider and linked them to the existing data in the data warehouse.
Previously, Niologic had combined multiple heterogenous data sources, including an SAP/S4systen, a VoIP telecom system, a shopping system, and further marketing channels.
Existing ETL processes were extended. Subsequently, all data was imported into a data lake in Google Cloud Storage and loaded into the BigQuery Data Warehouse with Google Cloud Composer. Using polyglot business keys, Niolotig combined data from multiple sources into a singular Data Warehouse. After that, the data was refined for machine learning and analytics within dara marts.
Results and customer value
Based on the solution, described in this success story Niologic extended the data lake and the data warehouse with data from the shipping provider incorporating them with existing data. For the first, it became possible to calculate product costs analytically and factually based on individual bookings – regardless of shipping method or service provider. Calculating direct costing also became possible. Additionally, the dashboard’s tracking overview based on derivations of floating averages allows the client to recognize impediments. With a time series analysis for shipping processes, the dashboard is now capable to predict overloaded service providers based on the shipping volume.