Building a BI dashboard that visualizes distributed data and supports decision-making
At one company, information related to sales, commercial activities, inventory, accounting, and other areas was spread across multiple systems and Excel files. Every time the team needed to check key figures, they had to manually compile the data. This not only made report preparation time-consuming, but also created discrepancies in the figures and update timing referenced by each department, making it easier for the speed and accuracy of decision-making to be affected.
SMILE was responsible for consolidating multiple data sources, designing KPIs, building the data aggregation platform, and designing and developing BI dashboards tailored to business management needs. This project sample showcases a BI dashboard solution designed to demonstrate how such a system can be implemented in a real operational environment.
- Focus Areas: BI / Analytics / Dashboard
- Challenge: Fragmented data, time-consuming Excel aggregation, and report creation dependent on specific individuals
- SMILE scope of support: Data source analysis, KPI / data model design, ETL, API, dashboards, filters, and access control
- Technology Stack: React, ECharts, Python FastAPI, PostgreSQL, AWS
- Timeline: 3 months
- Team Structure: 4 members / 10 MM
- Expected Scope: 3 data sources, 15 KPIs, and 1 management dashboard
- Expected Value: Reduced reporting workload, improved data transparency, and faster decision-making
Customer Challenges
Because the customer managed the data needed for operations across multiple systems and Excel files, they had to perform aggregation work each time they wanted to understand management conditions or on-site operations. When each person in charge uses a different aggregation method, even the same topic can be based on inconsistent meanings of figures or target periods, resulting in additional time spent on confirmation and adjustment.
When reporting relies heavily on manual work, update frequency declines, making it harder to check the latest situation when needed. In particular, when metrics spanning multiple departments such as sales, business development, inventory, and accounting cannot be viewed in one place, that lack of visibility itself can delay decision-making.
As a result, the need was not simply to add more charts, but to organize distributed data and create a mechanism that allows the indicators needed for management to be viewed according to the same standards.
Project Objectives
The objective of this project was to organize information spread across multiple data sources and build a dashboard environment that enables the required KPI to be monitored centrally.
At the same time, the goal was to reduce the burden of manual report creation and align the perspectives and aggregation standards that had differed by department, creating an information foundation that enables faster, easier decision-making.
SMILE’s Scope of Support
SMILE was responsible for reviewing the relevant data sources, designing KPIs, organizing the data model, building ETL jobs, implementing APIs, developing dashboard screens, designing filters, and implementing access control.
We go beyond simply building screens by clarifying which metrics should be viewed and for what purpose, then restructuring the information into a format that is easy for management and each department to use.
Key Development and Improvement Points
The key focus of this initiative was not simply to create an attractive dashboard, but to organize distributed data into a form that could be used for management. If source data is visualized while formats and update timings remain inconsistent, it becomes difficult for teams to use effectively in day-to-day operations.
SMILE therefore began by organizing the characteristics of each data source and clarifying which indicators should be managed using common standards. Based on this, we designed the definition of each KPI, the required aggregation units, viewing permissions, and filter conditions, creating a structure that enables users to quickly reach the information they need for their work.
We also designed the dashboard for use not only by a limited group of staff, but by users in different roles, with a strong focus on separating visible information according to permissions. This makes it easier to handle both the overall view needed for management decisions and the detailed information that frontline teams need to see.
Technologies Used
Frontend
We built a web dashboard with React and ECharts that supports KPI lists, chart rendering, date range selection, and filtering by conditions.
Backend
Using Python FastAPI, we implemented dashboard APIs and data integration processes, providing aggregated results from multiple sources in an easy-to-reference format.
Database
We used PostgreSQL to organize the data structures needed for KPI reference and store the data in a format that makes aggregation easier.
Cloud
Assuming AWS, we planned a configuration that makes dashboards, APIs, and scheduled aggregation easy to operate.
Other
We combined ETL / ELT, Excel import, and scheduled aggregation to create a configuration that leverages existing operational data and connects it to visualization.
Technical Challenges and How We Addressed Them
The challenge in this project was not simply displaying information from multiple data sources as-is, but organizing it into metrics that could be compared on a consistent basis. When the format or level of detail of the source data differs, even figures that appear similar can lead to incorrect decisions if the aggregation conditions are not aligned.
Because dashboard users need different information, another important consideration was how to separate high-level metrics for executives from detailed metrics for department managers. Too much information makes the dashboard difficult to use, while narrowing it down too far leaves gaps in day-to-day operations.
SMILE first organized the target data and KPI definitions, then clarified the required aggregation units and user perspectives before designing filters and permissions. This approach aims to create a dashboard structure that supports management decisions, rather than just a collection of charts.
Implementation Benefits
In this project, we reduced the burden of creating reports by switching between multiple files and systems, and created an environment where key indicators can be easily checked on a single screen. By automatically consolidating and visualizing data, dispersed information is easier to understand at a glance, helping reduce the effort required for report creation and improve the speed of decision-making.
Another major benefit is that by organizing KPI definitions and display criteria, interpretations of the numbers are less likely to vary by person, making it easier for teams across departments to discuss performance based on the same assumptions.
Key points for an offshore development structure
In BI dashboard projects, success depends not only on screen implementation, but also on defining metrics, understanding the data structure, and organizing viewing requirements for each user group. At SMILE, the teams in Japan and Vietnam work together to align on what should be displayed and at what level of detail it should be aggregated, then proceed with design and implementation.
By organizing requirements through BrSEs, sharing progress between Japan and Vietnam, conducting reviews, and aligning testing perspectives, SMILE makes it easier to move dashboard projects forward while minimizing gaps in understanding, even when requirements from multiple departments are involved.
Conclusion
The value of a BI dashboard is not in simply arranging charts, but in organizing dispersed data into a form that supports management decisions. In this case, SMILE handled the entire process, from reviewing data sources and designing KPIs to ETL, APIs, dashboard development, and access control, helping build an information foundation for decision-making.
For companies whose data is managed separately across multiple departments and that spend significant time preparing reports and understanding current conditions, developing this type of BI foundation is an effective option.
Looking for a reliable system development partner?
- At SMILE, we provide end-to-end support tailored to your business challenges and development needs, from requirements analysis and system design to development, testing, operation, and maintenance
- We support a wide range of projects, including new system development, enhancement of existing systems, legacy system modernization, AI/DX initiatives, and system
- By combining our Japan-based project coordination with our development team in Vietnam, we deliver flexible support for everything from small-scale enhancements to long-term system development and maintenance