Blog
Data Visualization in the Workplace
From raw data to a meaningful dashboard
Every company has data these days, but it is only through visualization that this data becomes a basis for decision-making. Yet not all companies derive the same benefit from it.
Through data visualization, complex raw data can be transformed into clear reporting portals, dashboards, and other analyses. They create transparency, support decision-making, and help optimize business processes.
In this blog post, we use a concrete example to show the steps required to lay the groundwork for meaningful visualizations and what these might look like in practice.
What is data visualization?
Data visualization refers to the graphical presentation of data and information so that it can be presented and interpreted in an appropriate form. Common forms of visualization include pie charts, bar charts, and trend lines. The use of such methods makes it possible to group data and prepare it for interpretation by specific target audiences. Data intended for visualization is typically structured data stored in a tabular format. A good example of this is standard Excel spreadsheets containing data from a CRM or ERP system.
Why visualize data?
A key benefit of visualizations is that people often grasp information more quickly through images and patterns than through text, databases, or tables. Furthermore, the right tools make it possible to tailor data visualizations specifically to a company’s specific questions and requirements.
For example, charts and graphs can show at a glance whether there are rising, falling, or unusual trends. Data visualization makes it easier to identify connections between isolated values, and key elements of datasets can be specifically highlighted. These features support both decision-making and the promotion of shared understanding.
Data Visualization Example – Gartner Data “Magic Quadrant”
A well-known example of data visualization is Gartner’s “Magic Quadrant” for analytics and business intelligence platforms. It illustrates the market positions of various vendors in this sector. Visualizations like this make it possible to identify market leaders, challengers, visionaries, and niche players at a glance.
The Path to Data Visualization
When does Excel stop being enough?
Smaller datasets can be easily visualized using charts in common business tools such as Microsoft 365 PowerPoint or Excel without much extra effort. However, this can prove more difficult with larger data sets (100,000 rows and more). Especially when using multiple data sources, it’s easy to lose track of the big picture. Accordingly, it’s important for companies to be able to assess at what point Excel charts no longer adequately meet the requirements for data visualization. Such an assessment isn’t always easy and requires expertise.
Specification – Identifying Business Use Cases for Visualization
To get started on larger visualization projects, it is necessary to assess the company’s requirements. Meaningful business use cases can then be derived from these requirements.
To identify such use cases, the following questions, among others, should be answered:
- What data is relevant?
- Where does the data come from?
- Which target groups will make use of the visualization?
- What decisions or processes should the visualization support?
These questions serve as an initial basis for further specification and can be expanded upon depending on the use case.
The level of detail in the specification may vary depending on the use case. The more relevant questions that are asked and answered, the more precise the specification will be.
Once the requirements and the associated business use cases have been defined, the technical implementation must be prepared. The primary focus of the technical implementation is to determine which data sources are needed and how they can be structured and prepared in a meaningful way.
Identify data sources
For the business use cases defined in the specification, the relevant data sources must first be identified. It is necessary to distinguish between different types of data sources, as they may vary in terms of origin, structure, and availability.
The number of data sources to be considered depends directly on the scope and complexity of the specification. In this process step, it is important to select the appropriate type of data source for each use case to ensure a meaningful data foundation.
Generally speaking, the following types of data sources can be distinguished:
| Data source type | Examples |
| Internal data sources | Company data, sales data, log files |
| External, public data sources | Destatis, Eurostat |
| External commercial data sources | Market research data, financial data providers |
| Open data platforms | GovData, Kaggle |
| Real-time & sensor data | IoT data, weather data, or traffic data |
| Survey – & Primary Data | Online surveys, interviews |
| Web & Social Media Data | Website tracking, social media data |
This blog post focuses on internal data sources, as these are readily available in many companies and offer a practical introduction to the data visualization process.
To illustrate the entire process using an example, we will use the following business use case:
“As a CRM manager, I want to know which customers we’ve completed the most projects for as a company.”
A company-wide sales pipeline is typically based on data from a CRM system. However, to identify relevant data sources, it is necessary to check whether there are dependencies on connected peripheral systems.
Such peripheral systems can include, for example, ERP systems, separate financial/accounting systems, or monitoring and marketing tools that contain supplementary information.
These dependencies must be taken into account to ensure a complete understanding of the data foundation and to avoid subsequent inconsistencies in the visualization process.
Data Staging
Once all relevant data sources have been evaluated and access has been secured, data staging can begin. Data staging encompasses all the extract, transform, and load activities necessary to prepare a suitable dataset for subsequent visualization.
These steps are collectively referred to as ETL and are central to modern data processing.
Extract
(Raw) data is extracted from an identified data source (sales data from a CRM system)
Transform
The extracted data is cleaned, standardized, and prepared. The complexity of this step depends on the condition of the raw data and the structure of the target system.
Load
The processed data is loaded into a target system, such as a data warehouse, a database, or other data structures that serve as the basis for data visualization.
In addition to the ETL approach, there is also the ELT approach, in which data is first loaded into the target system after extraction and then processed directly there.
This approach is used particularly in cloud data architectures.
The choice between ETL and ELT approaches depends on the specific business use cases and the underlying architecture. In many cases, implementation takes place within a DevOps-oriented infrastructure, where automated data pipelines extract raw data from the identified data sources and process it further.
Data pipelines are a central component of DevOps and are closely linked to continuous integration and continuous delivery (CI/CD) approaches, which we’ve explained in more detail in another blog post:
CI/CD and IaC
from Thomas Somogyi
Selecting the data storage device
During the “Load” step, the processed data is loaded into a data management system. As with the previous steps in the data visualization process, the choice of this system depends heavily on the business use case being implemented.
In addition to the business use case, requirements regarding manageability, cost, and performance are key factors that must be taken into account.
For smaller analyses, common collaboration tools such as Excel, SharePoint, or Access databases can be used. For larger and more complex projects, however, relational databases or data warehouses are employed.
While flexible tools enable fast and cost-effective implementation, structured data storage solutions offer a unified data foundation, better traceability, and scalability, which contributes to the sustainability of future projects.
Data Visualization – Implementation
Once all previous steps in the process—such as data preparation, data modeling, and the selection of an appropriate data platform—have been completed, the implementation of data visualization can begin.
As with the selection of a data storage solution, the type and format of data visualization depend on the business use cases specified. Different use cases require different types of visualization, such as operational dashboards for daily monitoring, analytical reports for business departments, or management dashboards for strategic decisions.
Another key aspect is defining the target audiences. Depending on the user group (management, data analysts, non-technical users), the requirements for usability, interactivity, and level of detail vary. While management-level users typically require aggregated metrics and clear visualizations, subject matter experts need more hands-on functionalities such as filtering, tabular views, etc.
Due to the varying requirements of the target groups, different data visualization tools are used depending on the use case and level of complexity. These differ in terms of feature set, usability, flexibility, and the technical expertise required.
The following table provides an overview of common visualization tools and their typical use cases:
| Tool | Examples of use |
| Tableau | Interactive dashboards for business units and management, exploratory analytics, self-service BI with a focus on visual data exploration |
| Power BI | Company-wide dashboards, management reports, integration with the Microsoft ecosystem (Excel, Azure, Teams), role-based access |
| Microsoft Excel | Simple reports, quick analyses, table- and pivot-based visualizations, prototyping |
| Sigma | Cloud-based analytics on data warehouses, tabular and visual reports |
| Python | Customized and highly flexible visualizations, exploratory data analysis, automated reports, and complex analyses using libraries |
In the following example, Power BI is used as a visualization tool.
Business Use Case: Implementation with Power BI
The procedure for implementing the blog use case is based on the content discussed so far:
As a CRM manager, I’d like to know which customers we’ve completed the most projects for as a company.
To keep this content brief, the level of detail in the specification has been kept to a minimum.
The following table provides a concise overview of the questions that have been answered for this visualization in the form of a “mini-specification”:
| What | Content |
| Data source | CRM system |
| Data volume | Small |
| Data Quality | Good |
| Data format | Semi-structured (XML export) |
| Target audience | CRM Manager (fictional role) |
| ETL or ELT? | ELT |
| EL Tool | Azure Data Factory |
| Data Storage (+ Transform) | SQL Database (Azure-based) |
| KPI | Number of client projects per client |
| Visualization tool | Power BI |
| Form of visualization | Bar chart |
| Filter | x-axis: Customer, y-axis: Number of customer projects |
In this example, the CRM system of a consulting firm serves as the data source. A script is used to export a defined subset of the relevant data in the form of an XML file. This XML file is then loaded into an Azure SQL database via a data pipeline. The SQL database serves as the central data repository for implementing the described use case and as the source for data visualization. Within the database, the raw data is transformed using SQL views: customer names are anonymized, and the number of projects per customer is aggregated.
The Azure SQL database is connected to Power BI via an integrated interface. Once connected, the data can be further modeled, transformed, and enriched with calculations within Power BI as needed. On this basis, the key metrics are calculated and displayed in the visualization for the CRM managers.
The final result of this process is a clear diagram:
Conclusions
The path to data visualization involves much more than just creating charts. It follows a context-dependent process that begins with defining business use cases, addresses data preparation, and extends all the way to selecting the right tool. As the process described above shows, careful specification forms the foundation of any successful visualization.
Those who ask the right questions, identify the appropriate data sources, and consistently follow the ETL or ELT process can create a reliable data foundation—and thus lay the groundwork for meaningful visualizations that support decision-making.
The business use case presented here using Power BI demonstrates that even simpler use cases still require a structured approach. The result illustrates how processed data can be presented in a way that is tailored to the target audience, provided that the infrastructure and the associated process are set up appropriately from the start. At the same time, the implementation of the use case highlights potential areas for improvement. For example, the XML export from the CRM system could be automated, and the Azure infrastructure could be further optimized in terms of costs.
Would you like to turn your data into clear, actionable visualizations?
We’ll support you every step of the way, from the initial use case idea to a meaningful dashboard.
