Personal projects
Personal projects
Insurance Fraud Analysis
As part of my journey as a Data Analyst, I developed an analysis based on a public Kaggle dataset containing 1,000 auto insurance claims recorded in the first quarter of 2015. The goal was to build an interactive Power BI report that explores three key dimensions: fraud detection, risk profiling, and claimant demographics. Using tools such as Excel, Power Query, and Power BI, I extracted and transformed the data, creating dynamic visualizations and simplified models consistent with the type of insurance analysis proposed. The report was structured to highlight fraud rates and suspicious collision types, damage severity and peak risk hours, as well as the geographic and occupational distribution of claimants. Each component was designed to encourage autonomous exploration, with dynamic filters and tooltips, within a clear and effective interface.
The full project is available at the following link: "Insurance claims".
As part of my final project in the Data Analyst program, I chose to conduct an in-depth analysis of the 2024/25 season of Como 1907, newly promoted to Italy’s Serie A.
I developed an interactive Power BI report aimed at analyzing the team’s performance from multiple angles: individual player contributions, tactical patterns, and season objectives tied to league survival. Using tools like Excel, Power Query, and Python, I extracted and transformed the data to create dynamic visualizations and simplified star schema models aligned with the nature of the sports-focused analysis.
The report is structured into thematic sections that cover goal analysis (scored/conceded every 15 minutes), key offensive and defensive patterns, stadium attendance comparisons between Serie A and Serie B, and the monthly breakdown of points needed to reach a realistic survival target.
Every component of the model is designed to promote autonomous exploration, featuring contextual tooltips and dynamic filters within a user-friendly and effective interface.
The full project is available at the following link: "Como 1907".
As part of an analysis based on the public dataset of the Brazilian e-commerce platform Olist, I developed an interactive Power BI report to examine online sales from 2016 to 2018.
The project involved restructuring the dataset into a star schema, integrating a calendar dimension, and creating a dynamic interface with filters, tooltips, and navigation buttons. The report includes sections dedicated to revenue and order trends, product distribution by category, customer behavior and review analysis, and the identification of anomalous orders.
The project is available at the following link: "Olist e-commerce".
I created a report using Looker Studio on the Sanremo Music Festival, covering the entire time span from 1951 to 2023.
The project involved analyzing the geographical origins of both hosts and winners to identify any recurring patterns or imbalances in territorial distribution over the years. It also explored the historical trends in TV audience and share, highlighting key moments of success or decline. To complete the analysis, data on the average age of winners and presenters for each edition was examined.
The project is available at the following link: "Festival di Sanremo".
As part of an analysis, I developed a report using Python in a Jupyter Notebook environment, based on the COVID-19 dataset curated by "Our World in Data".
The project involved an in-depth exploration of global health data, with both continental and national-level focus. For each continent, total recorded cases since the beginning of the pandemic were calculated, along with their percentage share of global cases, annual trends, and the progression of new cases. A boxplot was created to compare ICU patient trends in Italy, Germany, and France between May 2022 and April 2023. Additionally, the total number of hospitalized patients during 2021 was calculated and visualized—both numerically and graphically—for Italy, Germany, France, and Spain, in order to highlight potential differences between healthcare systems.
The project is available at the following link: "Covid19".
In this project, I designed and implemented a relational database using MySQL for ToysGroup, a company specializing in toy distribution.
Starting from domain analysis, I identified the main entities (Product, Region, and Sales) and defined the relationships and hierarchies among them, ensuring referential integrity and minimizing redundancy. The model was developed with a normalized structure and designed to support efficient sales analysis by geographic area and product category.
The project is available at the following link: "ToysGroup".
As part of a project for the Marche Region, I developed an interactive system for consulting and analyzing hospitality facilities across the regional territory.
The main goal was to create a dynamic Excel interface that allows users to select a specific facility via dropdown menus and automatically view all related information (city, address, email, etc.). Additionally, a second feature was implemented to display all facilities, filterable by city and category.
The project is available at the following link: "Strutture ricettive".
Team projects
As a team, we developed an interactive Power BI report based on the AdventureWorks database, with the goal of analyzing revenue, costs, and commercial performance. The project included data modeling, the creation of DAX measures, and the implementation of dynamic visualizations.
The report enables flexible, ad hoc exploration of sales data, highlighting the most dynamic product categories over time.
The project is available at the following link: "AdventureWorksDW".
During a group project focused on Looker Studio, we received a messy and unstructured dataset containing Oscar related information, provided directly by the client. The file, a .zip archive full of unverified CSVs, required extensive cleaning and transformation.
As a team, we worked on normalizing the data and restructuring it into coherent tables. We then designed an interactive report that offers an in-depth analysis of two key Academy Awards categories, highlighting trends, recurring patterns, and cross-cutting insights useful for understanding the dynamics behind cinematic success.
The project is available at the following link: "Oscar ".
During the Build Week, we conducted a comparative analysis using Python between Vans and Converse, focusing on pricing, product categories, and gender segmentation. After collecting and cleaning the data, we developed an interactive report with charts and tables to highlight positioning differences, brand strengths, and areas for improvement.
We structured the dataset by extracting key information from product names (such as model, category, and price) and normalized the data related to pricing and gender. The results provided valuable insights into the commercial strategies of both brands and supported data-driven decision-making in the retail sector.
The project is available at the following link:"Calzature".
During a group project, we designed and implemented a relational database using MySQL for VendiCose SpA, a supermarket chain, with the goal of efficiently managing the flow of orders between warehouses and retail stores. The system tracks sales in stores, automatically updates warehouse stock levels, and triggers restocking when a product falls below a predefined threshold, specific to each category and warehouse.
The project required careful data modeling and the management of many-to-many relationships between warehouses and stores, ensuring consistency and real-time traceability of operations.
The project is available at the following link:"VendicoseSPA".
To support the strategic decisions of the pharmaceutical company XYFARMA, we developed an interactive Excel report aimed at analyzing the trends of COVID-19 cases, recoveries, deaths, and vaccinations in Italy from 2020 to the present. After collecting official data, we cleaned and transformed it using Power Query and built a data model in Power Pivot for dynamic analysis.
The report includes pivot tables and charts, an interactive dashboard for querying data by region, and visualizations focused on the relationship between infections and vaccinations. The project provided XYFARMA with a solid foundation for evaluating potential investment in a new vaccine.
The project is available at the following link:"Covid-19".
© 2025 Fabrizio Murgo - All rights reserved