Mitsubishi Electric, a global comprehensive electrical equipment manufacturer, is transforming into a “Circular Digital-Engineering Company” that creates new value through data. To make that vision possible, the company launched the DX Innovation Center (DIC) as a company-wide project and built Serendie®, a digital ecosystem for technology, co-creation, human capital, and project promotion.
But a few challenges remained with previous data analysis: Analytics were fragmented, slow, and difficult to scale across teams. Manual workflows, disconnected tools, and inconsistent sharing of insights limited decision-making and prevented the company from applying analytics and GenAI at enterprise scale.
Also, for the Serendie data analysis platform, Mitsubishi Electric needed a data analysis environment that could support both data experts and non-experts while integrating directly with Snowflake and other data pools.
That’s when they turned to Dataiku, The Universal AI Platform™ for data prep, modeling, visualization, and app building.
Results With Dataiku
Dataiku is now a foundational component of Serendie (alongside Snowflake), enabling Mitsubishi Electric to expand the use of data analytics across the business.
Streamlined Delivery & Faster Time to Value
By consolidating data ingestion, preparation, modeling, visualization, and reporting into one unified platform, Mitsubishi Electric dramatically reduced effort and increased speed:
- 60% reduction in workload from data prep to reporting
- 80% reduction in time needed for visualizations compared to Python
- 40% reduction in time spent creating system documentation (i.e., functional specs, system design documents)
Cross-functional teams now collaborate directly inside shared Dataiku flows, allowing insights from engineers and analysts to be shared quickly within the project.
Scaling Expertise Beyond Specialists
With Dataiku’s visual, code-free workflows, non-technical domain experts can now participate in the data analysis process. Citizen data scientists leverage data analysis in their work, enabling them to make more informed decisions. Reader licenses and Dataiku applications help analysts communicate insights quickly, strengthening alignment across the organization.
Breakthrough Railway Insights in Just 2 Weeks
Using Dataiku, Mitsubishi Electric analyzed electricity usage, train operation data, and regenerative energy, achieving the following results toward decarbonization:
- Visualize surplus regenerative energy generated during braking.
- Propose solutions for utilizing surplus energy.
These insights, delivered in just two weeks, became worthy of industry attention and opened the door to new business opportunities and practical operational improvements.
Thermal Energy Optimization in Just 20 Days
Mitsubishi Electric worked to optimize energy demand for heating and cooling systems across multiple buildings. With Dataiku:
- Experts, analysts, and clients collaborated in a single, shared workspace.
- The team produced comprehensive visualizations and applied AutoML.
- Sophisticated thermal energy analysis and reporting for a full year of data was completed in just 20 days.
The Dataiku flow captured every step, enabling consistent documentation and fast iteration.
GenAI: Rag for Failure Response
Using Dataiku’s GenAI features, the DIC built a retrieval-augmented generation (RAG) system for failure response history.
The system incorporates text-based failure response records, PDF response manuals, replacement part information, and past response actions that make it possible to propose solutions, predict failures, and power multi-step chatbot interactions.
The Dataiku Difference
Mitsubishi Electric selected Dataiku because it uniquely supports both data expert and non-expert teams while enabling everything from data prep to app building. Key capabilities that enabled this success include:
- Unified workflows for data prep, modeling, documentation, and apps, all in one place
- No-code to full-code collaboration across engineers, analysts, and domain experts via visual workflows, shared dashboards, and governed sharing of projects and insights
- Flexible integration with Snowflake, cloud platforms, and the Serendie ecosystem
- GenAI with RAG, GenAI recipes, and Dataiku Answers
Vision & What’s Next
As Mitsubishi Electric expands the Serendie initiative, Dataiku provides the environment needed to scale analytics and AI across business units. Through this initiative, Mitsubishi Electric aims to deepen its analytical capabilities, accelerate GenAI efforts, and reach its goal of growing Serendie-related business to 1.1 trillion yen by 2030.
