Why I designed this Business Dashboard?

As the company grew, it became harder to track performance using spreadsheets, scattered reports, or conversations. We had two business areas — software development and IT education — each with different services, clients, and goals.

I saw an opportunity to create a central dashboard to help the team:

• Get a clear, real-time overview of business health

• Make faster, data-driven decisions

• Spot trends and issues early (e.g. drop in revenue, rising client churn)

• Save time on monthly reporting and reduce errors

• Support strategic planning across both business units

Many of the numbers we needed were already available on computers, such as in our finance system, booking tools, sales records, and client lists. That’s why it made sense to use existing digital data, clean the data, and make it visible in one place.

Where was all the data sourced from?

• Excel Files (CSV Reports): Manual exports from accounting tools and business logs

• Project management tools (Atlassian Jira and Confluence): Exported task, project, and time reports (CSV, JSON). 

• Invoicing and Financial Data (szamlazz.hu): Downloaded invoice records as structured files, cleaned and formatted them, and integrated into relevant financial tables (e.g.sale)

• Website Traffic (Google Analytics): I connected with the API to collect user activity metrics and processed them to gain insights into engagement related to services and training.

• Training Registration Data: Manual Excel entries and website form submissions

• Other Documents and Files: other structured files were processed and joined with core tables using identifiers like client_id, service_id.

How did I use APIs

To integrate website traffic data into my dashboard, I connected to the Google Analytics Data API (GA4) using Python. Once the data was returned in JSON format, I converted it into a structured Pandas DataFrame, cleaned the column names, and transformed the metrics into the correct data types. I then loaded this cleaned data into my PostgreSQL database, where it could be joined with other business data and visualised in Tableau.

For the Python code, please visit my GitHub site: My Projects on GitHub

How I Organised the Data

I created the table plans by designing an Entity-Relationship Diagram (ERD) in LucidChart. First, I identified the key business entities based on the KPIs I wanted to measure, such as clients, services, sales, and training sessions. Then I defined the relationships between these tables and added primary and foreign keys to ensure a logical data structure. Each table was planned with relevant fields and data types to support analysis in PostgreSQL. This diagram served as the blueprint for building a clean, efficient, and scalable relational database.

How I Analysed the Data

1. Structured the Data in PostgreSQL
After importing the data into PostgreSQL, I performed:

1. • Normalisation and table relationships (using foreign keys)

   • Handling missing and inconsistent values

   • Date conversions and data type alignment

   • Calculated fields such as profit margin, churn rate, and session attendance

2. Exploratory Data Analysis (EDA)
I wrote SQL queries to:

1. • Summarise trends (e.g., revenue by month, sessions delivered)

   • Filter and group data by categories (e.g., service type, client segment)

   • Detect outliers and validate assumptions

3. KPI Calculations and Metric Design
I created queries for each KPI, ensuring the logic was consistent with business goals (e.g., calculating Monthly Recurring Revenue (MRR), Net Income, and Task Automation Savings).

4. Dashboard Design in Tableau
I visualised insights using:

1. • Bar charts, line charts, and KPIs for trends

   • Filters and parameters for interactivity

   • Storytelling elements to guide business users through the findings

5. Validation with Stakeholders
I presented early dashboard versions, collected feedback, and iterated to ensure the results answered the right questions and supported decision-making.

The business question was: Which services are the most profitable?

The KPIs

• Revenue per Service

• Cost per Service

• Profit = Revenue – Cost

• Profit Margin (%) = Profit / Revenue

 SQL Queries
 

My answer to this question is as follows:

The Junior Java Programming Course clearly stood out, with a 66.67% profit margin and the highest overall revenue (76.8M HUF). This indicates strong market demand and efficient cost management. Business Coaching also performed very well, achieving a 62.50% margin, despite being a lower-volume service. Other services, such as Frontend Bootcamp and Code Audits, were still profitable but with lower efficiency. 

My recommendations based on the results are for example:

1. Scale the Junior Java Programming Course: It combines high demand with excellent margins. Increasing the number of sessions, offering advanced versions, or building corporate packages could amplify both reach and revenue.

    

2. Promote Business Coaching as a Premium Offer: With a strong margin and lower delivery costs, this service is a high-value, low-risk option, and targeting executive clients or bundling with training could add strategic value.

Future Results

Following my analysis, the company implemented my recommendation to scale the Junior Java Programming Course. Within the year, they launched corporate packages and introduced advanced and senior-level versions of the course. These new offerings were well-received, significantly increasing revenue from training services and strengthening long-term client relationships.

For more queries, please visit my GitHub site: My Projects on GitHub

Defined the Dashboard Layout

To create a clear and insightful dashboard, I followed a structured design process that combined UX principles, business logic, and visual storytelling.

1. Sketches

I began by sketching the layout, using tools like Miro to visually map each business question to its corresponding KPIs. This helped ensure that the dashboard was focused on real strategic needs rather than just data display. Each KPI was tied directly to a business goal—such as tracking profit margins, client churn, or training impact—which shaped the structure and flow of the dashboard.

2. Design

Once the logic and content were mapped, I used Figma to design wireframes and mockups. This step allowed me to experiment with layout ideas and structure the content in a way that would be intuitive for decision-makers. I applied my UX/UI experience to prioritise visual clarity, ensuring that the most important KPIs appeared at the top, with supporting metrics grouped logically below them. This top-down design supports quick scanning and action-oriented insight.

With the layout defined, I strategically selected visual types that best matched the nature of the data:

• Line charts were used to show trends over time, such as revenue growth or client acquisition, offering a clear view of business progress.

• Bar charts were ideal for comparisons, such as analysing profit by service type.

• Scorecards and KPI boxes were placed prominently to highlight single key figures, like churn rate or number of active clients.

• Tables were used to present detailed, sortable lists—such as feedback entries or upcoming training sessions—providing context where needed.

3. Tableau

The dashboard was then developed in Tableau, where I connected it directly to the PostgreSQL database that held the cleaned and structured data. I created calculated fields for metrics like profit margin, growth rate, or time saved, and added filters and parameters to allow users to explore different segments, services, or time periods.

4. Iterations

Finally, I iterated the dashboard based on stakeholder feedback and visual clarity tests. This included adjusting chart types, refining layout spacing, simplifying labels, and ensuring all elements supported the original business questions. The result is a dashboard that is not only visually engaging, but also deeply tied to business goals, offering real-time insight and supporting strategic decisions across departments.

Publish and Presentation

After completing the dashboard design and analysis, I focused on presenting the results in a clear and engaging way for both technical and non-technical audiences. The project was also version-controlled using GitHub, where I documented the ER diagram, scripts, data cleaning and analysing steps.

VIDEO Demo Presentation

In this short video, I present the business analysis dashboard I designed and built, explaining the goals, data sources, key KPIs, and how the final Tableau dashboard supports decision-making. For the video, please visit my YouTube Channel: My Projects on YouTube

Disclaimer

This project presentation uses mock data to respect the confidentiality of the company I worked with. Some elements of the business story and structure have been modified or simplified at the company’s request. However, all content shown on this page has been reviewed and approved for public sharing. The analysis, structure, and dashboard design reflect my real work and problem-solving process.