Case | Automotive
Shared Data Platform for a Multinational Automotive Group
Dealer-system data from the Taiwan branch moves to the cloud automatically every day, giving overseas HQ real-time visibility into data quality and completeness.
Home / Case Studies / Multinational Shared Data Platform
Dealer-system data from the Taiwan branch was only used locally, so overseas HQ had no view of its quality. We defined the data to exchange and its quality rules, then built a three-layer data lake on Azure Synapse so local data moves to the cloud every day and HQ can track completeness and quality trends at any time.
| Item | Details |
|---|---|
| Industry | Automotive distribution & dealerships |
| Services | Data Platform & Data Engineering, data-quality monitoring |
| Technology | Azure Synapse, Azure Data Lake Storage Gen2, Power Automate, SharePoint, Power BI |
| Results | Five files in the cloud daily; quality issues found in 1 day instead of weeks; 30+ automated checks |
Background & challenges
- The dealer system produced five daily files—customers, vehicles, sales and more—used only in Taiwan and invisible to HQ.
- Overseas HQ had no real-time view of data quality or completeness, so problems surfaced only at analysis time.
- Cloud environments and accounts were managed centrally by HQ, requiring cross-border collaboration and compliance with HQ’s security rules.
Our approach
Following Datas’ four-step method: get the data right first, then deliver the application, then keep it running.
- Data audit: Audited the five daily dealer-system files and defined fields, data-quality rules and the metrics HQ needs to track.
- Build the platform: Power Automate delivers files daily to HQ’s SharePoint landing zone; Synapse Pipelines ingest them on a schedule into Raw / Staging / Curated layers in Azure Data Lake Storage Gen2.
- Reporting: Built data-quality fact tables and a star schema with Synapse serverless SQL, plus two Power BI report sets: Summary and Insight.
- Operate & optimize: Automatic alerts on pipeline failures; permissions managed centrally through HQ’s Entra ID with row-level security and sensitive-field masking; pay-as-you-go serverless resources keep costs low.
Results
- Taiwan data is centralized on the HQ platform automatically every day, with no manual transfers.
- HQ tracks completeness and quality trends with 30+ automated checks, cutting the time to spot issues from weeks to the next day so they can be fixed at the source.
- A low-cost serverless and data-lake architecture became a template that can be replicated in other markets.
- Unified permissions and sensitive-field masking meet cross-border data governance requirements.
| Benefit | Before | After |
|---|---|---|
| HQ data visibility | None (data used only in Taiwan) | In the cloud daily (T+1) |
| Finding quality issues | Only at analysis time, often weeks later | Checked daily, spotted the next day |
| Quality checks | No systematic checks | 30+ automated rules |
| Rolling out to new markets | Architecture planned from scratch | Template reused in about 2–3 weeks |
Related services & next steps
- Data Platform & Data Engineering: the core of this project—three-layer data lake, automated pipelines and data-quality monitoring.
- AI Agent Process Automation: can extend to automatic alerts and correction workflows for quality issues.
- AI Assistant: can let HQ ask about each market’s data in natural language.
More case studies
Aviation | Data Platform
Management Reporting Automation for an International Airline
Manual daily Excel reports replaced by automatic T+1 updates.
Foundation | AI Application
Cloud Data Platform & AI Data Query
Six data types integrated, with natural-language queries.
Food Retail | Predictive Models
AI Order Recommendations for Food Retail Stores
87.77% forecast accuracy, driving store order recommendations.
Wondering what this could look like in your industry?
Book a consultation and we’ll share more architecture details.