Swedish multinational engineering company Sandvik partnered with Xomnia to develop a reporting suite that generates better insights and ensures that their global fleet used in mining and construction operations is supported with parts and services.*
Strengthened by Xomnia’s analytics engineering expertise, Sandvik achieved a new reporting dashboard that is specifically designed to quickly get meaningful insights. This is done while still providing the option to access data repositories for those seeking a self-service approach.
The robust dashboard supports different stakeholders within the client while serving as a strong launching pad for future solutions too.
“When we onboarded Xomnia, we had many unhappy users as they were missing functionality. We have been able since to address many of those issues, significantly reducing our development backlog but also adding in many enhancements. Fleet Analytics is now much more stable and is a platform that we can continue to build on.” – Niek van Zijl, Head of global pricing & commercial analytics at Sandvik
Challenge
Sandvik wanted to transition to a new version of their fleet reporting tool, Fleet Analytics. The new layout was designed by another company, but the implementation was to be handled by the existing Sandvik fleet team. Meanwhile, Sandvik fully redeveloped their databases to a modern cloud infrastructure.
The migration to a new tool and the transition to all-new data sources, however, caused several problems with their existing user-base, both small and big. Confidence in the data that they were consuming was low, and required features were missing. At the same time, their development backlog kept growing quickly.
To help Sandvik meet the many requirements to transition to a new and enhanced version of the Fleet Analytics reporting tool, our Analytics Engineer Luis Lucchini joined them to support their development process.
Solution
Luis worked with a team consisting of Power BI and backend developers. Upon joining the client’s team, he focused on understanding the different stakeholders and users of the Fleet Analytics dashboarding solution, as well as their needs and user stories. Wearing the hats of different stakeholders, he tested different possible reporting solutions to choose the best one.
“There are several types of users for fleet analytics at Sandvik, most of whom just need some quick insights to make business decisions,” Luis explains. “Others are very data-driven and need to access very large amounts of data that can be overwhelming for almost everyone else.”
Following the initial research phase, our analytics engineer helped the client create a solution that involved six main parts:
- A semantic model consisting of a combination of various related tables structured in a star-schema.
- Data measures that Luis modernized and optimized. The measures define the aggregations on the dataset, and are painstakingly reviewed and revised to reflect the correct approach that the reports and dashboards will display. Their explainability and documentation must always be paramount in the development process.
- A robust dashboard that involves several different report pages and sub-pages, each of which coming with a purposeful vision, layout, goal and set of metrics to satisfy user stories in a cohesive manner.
- A report layout that is designed to guide storytelling, but also provide a self-service hub for the users to build their own dataset, leveraging the new model and measures.
- A data quality component that allows stakeholders to go back to their source systems and know, for instance, the equipment that lacks certain information. This allows the users to fill in this missing information.
- Version control, which enables the analytics team to collaborate and expand to several new members without causing any issues. Version control also means that even if new data rolls out with errors, the pipeline can be stopped before there is significant impact to the users. Moreover, even if bad updates go through, the team can easily transition back to an earlier working version and keep things running smoothly with minimal impact to end-users.
The solution is supported by Azure cloud and is based on Databricks in the back-end, enabling the Data Mesh approach at Sandvik. Power BI sits on top for the reporting, and Azure DevOps for infrastructure and development.
Impact
Xomnia strengthened Sandvik with the eye for extracting insights out of data, and making reporting more impactful and more meaningful. The resulting reporting infrastructure is in production and has new releases every month. The data model created follows the best practices and cutting-edge technology integrations from Microsoft. This product is perfectly positioned to be developed further in various avenues quickly.
The development process has been sped up while the technical debt has decreased significantly. The time to deliver new reporting products has decreased, enabling several teams to enhance their customer interactions with data-driven insights. Additionally, confidence in the data has also been satisfied because of the extensive documentation provided and the clear communication of changes and updates.
* This fleet is used in mining and construction operations worldwide, which the aftermarket teams are aiming to keep running smoothly with minimal down-times. The Parts and Services division is responsible to proactively address potential issues in the mining and construction fleet.