Databricks Data Rooms allows self-service access to IoT data for Normet
Normet is a world-leading, innovative technology company defining the future of underground operations in mining, tunnelling, and civil engineering projects. Their work helps partners increase safety, sustainability, and productivity throughout the lifecycle of their projects.
Machines connected with a data and AI back end allow Normet to improve customer service, safety and utilize their IoT data more widely.
In a spearhead project of Normet, they seek to create an AI solution that gives stakeholders easy access to machine data - streaming from Normet machinery around the world to the Normet Data Platform. To achieve this new AI solution, Normet partnered up with Brightly to create a proof-of-concept solution using Databricks Data Rooms.
“We were amazed and thrilled about the out-of-box functionality of Databricks AI Data Rooms, as well as how quickly and efficiently we carried out the deployment with Brightly.“ –Pekka Jaarinen, Normet
Self-service access to IoT data, made easy
Together with Brightly, Normet connected Databricks Data Rooms with their Databricks lakehouse. Data Rooms is an AI-powered data exploration and generative AI platform that empowers technical and non-technical users to analyze company data in a Databricks lakehouse. Users ask natural language questions and the Data Rooms AI plans and executes the necessary queries in Databricks to answer the user’s questions in natural language, enriched with visualizations.
Data Rooms is designed so that no data from Databricks is sent to the AI model. In fact, only metadata, descriptions of the tables and columns, are supplied for the AI as it plans its analyses. This ensures absolute data security.
Here are some examples of questions the tool can answer:
1. What is the age and usage hours of the customer X’s machines?
2. Has the temperature of device Y’s compressor exceeded 50 in the last month?
3. Does any equipment show trend in increasing temperature for hydraulic oil for past month?
Out-of-the-box benefits to Normet
The proof of concept was conducted to verify that the Data Rooms AI system correctly answered pre-defined business and technical questions based on the connected machines and customer data. We tested the concept using different User Personas with questions of varying complexities, and – with minor Instructions to AI – Data Rooms answered correctly based on available data. With some questions, the user needed to further clarify initially ambiguous questions to receive correct answers. Visualization of the results as, e.g., charts and graphs worked out-of-the-box without any further adjustments.
As a proof of concept, the AI natural language query project was exceptionally successful, especially considering the out-of-the box value achieved. The tool demonstrated the ability to accurately answer complex business and technical questions, enabling Normet to develop solutions for non-technical users to effortlessly access and utilize complex IoT data from Normet’s connected fleet.
Towards more accessible and usable IoT data
Databricks Data Rooms demonstrated its potential as a solution for the needs of Normet making it easier for also non-technical users to access and utilize company data in a natural language format. Looking ahead, Normet aims to refine this solution to meet specific user needs, ensuring it is tailored to real-world applications. Normet will continue to make data more accessible and usable, empowering different stakeholders to make informed decisions and drive further improvements in their operations.
