Background and Challenge
Our client, Reveel, operates a professional shipping and logistics intelligence application that uses machine learning to perform auditing, reporting, and analytics on complex shipping data, identifying simple ways for their clients to save money.
They collect billions of rows of shipping data, including customer invoices and transactional data from UPS and FedEx, and they needed a way to organize it to increase the speed of their advanced analytics.
Their shipping data needs to be leveraged by multiple teams within the organization – including for their front-end user application to identify cost savings and for their data science team to build machine learning models. Thus, they needed to increase the scalability of their system to allow for multiple users and create a single source of truth database to standardize data for multiple use cases.
Reveel was founded by former DHL sales executives, with the mission to increase clients’ profitability by empowering them with shipping intelligence and advocating for transparency within the industry.
They help organizations ship smarter with easy-to-use tools powered by AI technology and have saved millions of dollars for their customers since their founding in 2006. Learn more about Reveel.
Reveel was searching for a strong Databricks partner to build their customer application and reached out to Adastra.
Adastra's dedicated team of data engineers specialize in Databricks and the AWS platform. We have extensive experience implementing scalable and robust analytics solutions on the cloud using these technologies. In addition, Adastra has built lucrative solutions in the supply chain and logistics vertical for our retail, manufacturing, mining, forestry, and natural resources customers.
Adastra builds data modernization solutions by harnessing our skills in data engineering and AI/ML modeling capabilities, making us the ideal fit for the project.
Databricks on AWS
Adastra began by reviewing the current application backend processes and proposed improvements to increase the throughput of data.
We helped the client store the transactional and shipping data on AWS, and then standardized and processed the billions of rows they are collecting using Databricks large-scale compute services.
Adastra then wrote logic in Spark to service data processed in Databricks to the front-end, customer-facing application. This allows Reveel customers to run large-scale analytical interactive dashboards to identify savings with a sub-5 second response.
Building a Single Source of Truth with Delta Lakes
Reveel needed to build a single source of truth database that could be queried for multiple use cases, without facing performance issues.
To achieve this, Adastra worked with multiple stakeholders in the organization to build low-latency, high-performance databases that connect to a single source of truth, while keeping the data standardized. We did so by leveraging delta lakes, which combine features of both data warehouses and data lakes into a single tool.
Adastra then helped Reveel set up Spark clusters in Databricks for their data science team to execute machine learning workloads for customer analytics.
The solution resulted in a standardized, scalable system that allowed shipping data to be leveraged by both the data science team, as well as for their customer-facing application.
Adastra’s solution made it possible for Reveel to increase the speed of processing for their customer-facing application as well as exponentially increase analytical performance and reporting. Where pulling analytics used to take 5-10 minutes with the old system, they can be pulled in under 5 seconds.
The system went from being able to handle 1-2 users at a time to up to 50 and Reveel’s data science team can now utilize data readily without their system being overloaded.
Adastra’s solution left Reveel with standardized golden records of high-quality data, and they are now able to onboard a larger group of customers to their shipping intelligence application.
Looking to learn more on how Adastra can help? Speak to one of our experts:
We will contact you as soon as possible.