i-5O Operationalizes Cloud Data Workflows in Hours with Self-Service ETL++

i-5O offers an AI insights platform that detects anomalies in automated production lines. Image data is streamed in real-time from a production line, where the company’s artificial intelligence software serves as the industrial engineer’s set of eyes. i-5O’s goal is to enhance human performance and safety within manufacturing by gaining unprecedented visibility into manual production processes.

As i-5O grows, the requirements for their data pipelines are changing. With this growth, it was essential to minimize the ETL process’s impact on source systems. i-5O needed a way to quickly and efficiently enrich complicated data pipelines to advance their machine learning models without disruptions. The team had to figure out a path forward and initially considered hiring a team of data engineers focused on DataOps, which would have taken months to staff and diverted critical resources away from strategic business initiatives.

A team of data scientists must format i-5O’s complex image data to precise specifications. These data transformations enable teams to model real-time data as it streams from the production lines. Effective implementation of the company’s machine learning models presented vital challenges, including:

• Building flexible data pipelines that adapt to unknowns in always-changing data models.
• Finding the best tools to integrate with the Microsoft stack for scalable processing of real-time video data.
• Enabling a rapid, self-service approach to building new data pipelines and the ability to automate the ETL process across cloud data warehouses.
• Enterprise-level support to assist with optimizing and scaling data in the cloud.

After researching options, i-5O decided to leverage Azure to implement scalable machine learning processes and analytics in the cloud and selected Datameer as their Cloud ETL tool to feed data to Azure Synapse. Datameer delivered with:

• Integration of cloud data assets — i-5O can now integrate structured, semi-structured, and unstructured data into Azure – to efficiently run complex models in the cloud.
• Self-service access to data — Datameer gave the i-5O data science team a rich set of tools to speed analytic cycles without staffing a team of data engineers.
• Operationalizing insights — the i-5O team securely prepares complex imaging data in Datameer  for automated integration with their proprietary machine learning models.
• Integrated analytics — Datameer’s spreadsheet-style interface, powerful functions, and workbook linking enabled i-5O to consolidate their analytics into one user-friendly tool.

Datameer liberated the data science team from the resource-consuming tasks of building data pipelines. With Datameer, the team is focused on growing i-5O’s business capabilities by building and refining machine learning models. Thanks to their decision to partner with Datameer, i-5O was able to:

• Rapidly prepare and integrate complex, real-time datasets for modeling across a variety of data sources.
• Months of coding saved per year by reducing the time it would have taken to hand-code complex data integration workflows from months to hours.
• Reduce the number of tools necessary. Datameer ETL++ includes a no-code pipeline tool (with 200+ connectors to cloud and on-premise sources) and a no-code data transformation tool – in one solution.
• Reduce the number of DataOps resources needed down to zero. The analytics and data science staff build ETLs using Datameer’s no-code interface.
• 10x faster analytics, boosting the productivity of analytics teams and giving them bandwidth for more strategic initiatives.