- Databricks Lakehouse for Retail is an integrated platform to help retailers tackle specific industry challenges, such as demand forecasting and supply chain inventory provision.
- The product helps users gain insights from data, either through traditional analytics or by leveraging Databricks’ AI tools.
Advanced analytics firm Databricks, known for its AutoML capabilities, has launched its first industry-specific product, Databricks Lakehouse for Retail. It is a savvy move to target the retail sector, which has been shaken to the core during the global pandemic, with supply chain disturbances and a massive shift towards e-commerce accelerating trends that would have taken many years in normal circumstances. Against this backdrop, retailers are looking to optimize their data to better manage inventory issues and supply chains and to improve their ability to forecast demand and personalize marketing campaigns for customers.
Databricks Lakehouse for Retail helps firms access and query data using tools designed to address their specific requirements. It features ‘accelerators,’ or blueprints containing best practices to solve business problems that are fine-tuned over time with the platform’s AI and ML capabilities. The goal is to shorten the time and effort involved to create the data models and associated analyses to support specific business processes.
Databricks seeks to address common challenges faced by large enterprises struggling to extract value from vast amounts of data. Its Databricks Lakehouse architecture, based on the open source MLflow platform and Apache Spark software, breaks down data silos by acting as a single repository that combines traditional data warehouses (with structured data) and data lakes (with unstructured data) in a unified platform (the Databricks Lakehouse).
The Databricks Lakehouse automates the complexity of building and maintaining pipelines and running extract, transform, and load (ETL) workloads. In the past, data scientists were required to make a copy of the unstructured data so that it could be structured and analyzed in a separate environment, but now it is possible to analyze all types of data in this unified platform. According to Databricks, more than 5,000 organizations use the technology to enable this data engineering at scale.
Databricks was born out of the drive to ‘democratize data and AI,’ by putting artificial intelligence (AI) and machine learning (ML) capabilities in the hands of ‘citizen data scientists’ with less technical expertise than traditional data scientists and data engineers. The company says it intends to ‘leave code behind,’ and to this effect, it acquired 8080 Labs, creator of UI-based data science tool bamboolib, last year. The bamboolib tool generates production-ready code to enable so-called ‘citizen data scientists’ to operate in a low-code environment, a capability that is now part of the Lakehouse architecture. The integration of bamboolib with AutoML technology allows users with only a basic understanding of data science to train machine learning models that optimize their company’s datasets.
Databricks’ competitors have already launched numerous industry-focused solutions, including for retail. DataRobot released its own vertical solutions, AI Cloud for Industries, with products for manufacturing, banking, healthcare, and retail. Google Cloud also has industry offerings targeting most vertical industries and, specifically for retail, applications such as Looker and Anthos, among others. Likewise, AWS has launched ML/AI solutions including Amazon HealthLake (healthcare), Amazon FinSpace (finance), and Amazon LookOut for Equipment (industrial machinery). The landscape of vertically oriented AI applications is getting crowded, and Databricks, which has been a trailblazer in the AutoML space, would be wise to keep pace with its rivals by updating its portfolio in an increasingly specialized marketplace where enterprises demand the right tools to best compete in their sector.