The main objective of this personal portfolio is to demonstrate my skills in solving business challengs through my knowledge and tools of Data Science.
Data Scientist
I have been working on IT projects in a relevant tourism and entertainment group for almost 11 years helping the company achieve its strategic objectives.
I have knowledge of all stages of developing a business solution using the concepts and tools of Data Science, from understanding the business to publishing the model in production using Clouds.
I have already developed solutions for important business problemas such as detecting fraud transactions, sales forecast and prioritizing customers for cross-selling.
The details of each solution are described in the projects below.
Analytical Tools: Data Collect and Storage: Spark(pyspark), Azure Data Factory, Pentaho Data Integration, Talend data Integration, SQL Databases: Databricks Delta Lake, MySQL, Postgres, SQL Server, Oracle Big Data: Spark Databricks Data Processing and Analysis: Python Development: Git, Scrum, Python. Operational System: Windows, Linux Data Vizualization: Power BI, Databricks SQL Machine Learning Modeling: Classification, Regression, Clusterization Machine Learning Deployment: Heroku Cloud Application Platform, Flask API
Links:
Machine Learning and predict next 6 week sales. Company need to identify the sales forecast to next 6 weeks to invest on shops reform. In this project, i built a Machine Learning regressort to predict sales forecast with 98,1% of accuracy. The model is available in telegram bot. Send a shop number (99) and he will responde you the sales forecast.
Machine Learning and taking fraud detection. Companies are reducing their costs with detecting fraudulent transactions, while companies providing theses types of services are increasing thier income. In this project, I built a Machine Learning classifier to label fraudulent transactions with 99.51% of accuracy. The performance of this model would bring revenue of U$623,2 millions according to the company's business model described in the problem definition.
Machine Learning and ranking customers. Company need to prioritizing 20000 customers to call offering a vehicle's Insurance. The model will prioritizing the more probality interesting of acquirin the vehicle insurance of all customer to maximize sales. The performance of this model bringed 298% increasing than random method, with revenue of U$5,8 millions according to the company's business model described in the problem definition. The model API is available in Heroku Platform
Machine Learning and predict cardiac diseases. Company need to increase accuracy on cardio detection, current doing by human specialists. The performance of this model bringed 7x increasing than current method, with revenue of U$41,9 millions according to the company's business model described in the problem definition.