Vinícius Borges
Data Scientist & Data Analyst. in Remote, He/Him
About
With over three years of experience in Data Science, I have built end-to-end data projects, mentored aspiring analysts, and collaborated with industry leaders, all with the intent of creating solutions that are strategically relevant and visually appealing. I sculpt the company’s data environment, ensuring it resonates across departments and aligns with the company’s ethos.
My passion is to create excellence in data projects from the ground up, from inception to raising the bar for quality in all aspects of the process. I also lead with a philosophy of endorsing continuous learning, distributing my knowledge to shape better talent. Believing that the right data solves any problem, I co-lead the Revenue Protection (Recovery) data team at Cemig, the world's most notable Brazilian PPP electric utility.
I have also worked with clients in Latin America over the past few years from a wide industry range, such as startups, fintechs, edtechs, food industry, marketing, Ads & among others.
Stack: Python, SAS, SQL, Java, Machine Learning, mySQL, Databricks, Apache Airflow, FastAPI, Docker, PostgreSQL, Mage, MS PowerBI, GCP, SAP, AWS, Pentaho, MS PowerBI, MS Sharepoint, Metabase.
"I want my projects to convey a message, presenting situations where the answer is not a clear yes or no." — Hideo Kojima.
Contact
Projects
In this project, I analyzed sales data, overcame customer objections and improved Red Bull's visibility in bars. Used DataViz, EDA and predictive models to increase assertiveness and sales performance.
Tools used include Excel, PowerBI, Plotly.js and Figma.
A Colab notebook that provides a training, validation, and prediction classification model for users who are likely to churn from their streaming service.
Writing
This article seeks a solution to reduce costs with traditional collection letters and increase revenue recovery.
The cost reduction of approximately 94% in sending letters allowed the collection of a greater number of base customers, considered a “suppressed charge”. In this context, the perceived increase in the amount collected through this strategy was R$22 million per month during the period analyzed, guaranteeing a greater recovery in revenue.
Work Experience
As a data analyst at Cemig, I create end-to-end data projects involving data with a focus on revenue recovery for the company. I also co-lead a data team to ensure our business dashboards and reports are accurate and up-to-date. Our stack is based on SAS, SAP CCS, Python, AWS, SQL, MS PowerBI, MS Sharepoint, MS Fabric and Azure Data Lake.
-
Tracked and identified fraud in the Electrical Sector by monitoring irregularities in Cemig installations' ownership transfers. Detected a debt surpassing US$65 million and mitigated it with our collection tools, employing geo-referential models for statewide fraud detection & implemented a delta visual tool for user-friendly analysis;
-
Developed a friendly algorithm to evaluate service effectiveness, improving accuracy by 84%, and created a ranking system used across all Cemig operational guidelines;
-
Remastered 29 data projects, and proposed restructuring to ensure stability and accountability in the sector's data;
-
Achieved 82% accuracy in forecasting material costs and implemented visualization techniques to increase collections by over US$992 million annually in CEMIG's disconnection services, across urban and rural areas;
-
Tools and technologies: SAS, SAP CCS, Python, AWS S3, AWS Athena, SQL, MS PowerBI, MS Sharepoint, MS Fabric, Azure Data Lake.