I currently work for a startup from Oslo, applying analytics and predictive modelling skills into improving the offering to clients on the platform. Prior to that, I worked as the Lead Optimization Analyst at a technology startup, where I performed site audits and used customer data insights to develop A/B and multivariate testing strategies for e-commerce websites to improve user experience and increase revenue; My previous experience includes working for a Brazilian Federal Government institution in their Budjet and Finance department; and for a metals and mining corporation alongside a multidisciplinary team identifying opportunities to improve productivity.
Data Analytics and Visualization Bootcamp
B.S. Civil Engineering
Smart Energy Systems
Excel 2016 Expert Certification
Python/Flask web application deployed to Heroku, developed as part of a team project. It's integrated with a PostgreSQL database, accessed using SQLAlchemy, and a Machine Learning model to make predictions based on user input, which is processed using NLP. Additionally, Plotly visualizations provide further insights into the data.
The project aims to answer the question of whether a song's lyrics can predict its success, which was defined as reaching the Billboard Hot 100 list, and exploring other possible factors that can contribute to that success.
My part in the project was the development of the web app, both client and server sides; the integration with the Machine Learning model and the database (hosted in AWS), as well as the NLP to process the user data to be fed into the model; the Plotly visualizations and the deployment to Heroku.View in GitHub View Deployed App
The project consisted of creating a proposal for an angel investor in a fictitious bike-share business located in the city of Des Moines, Iowa, similar to the CitiBike program in New York City.
The analysis includes details on whether the business would work based on how it works in New York City and a Tableau story with insights extracted from the data.View in GitHub
The purpose of this project is to visually show the differences between the magnitudes of earthquakes all over the world for the last seven days. To illustrate the severity of earthquakes in relation to tectonic plates, an API call to the tectonic plate data is made using D3, and the data is added as an overlay to the map using the GeoJSON layer.
The project consisted of using Python and Jupyter Notebook to evaluate different models with unbalanced classes, including using ensemble and resampling techniques, to determine whether they should be used to predict credit risk.
Logistic Regression, Decision Tree, Random Forest and Support Vector Machine algorithms were implemented and compared.View in GitHub
This project consisted of performing statistical tests to create a technical report for the AutosRUs' MechaCar fictitious product development team to justify design choices.
The analysis involved data cleaning and reshaping using tidyverse in R; visualizations plotted using ggplot2; implementation and evaluation of simple and multiple linear regression models, one-sample t-Tests, two-sample t-Tests, Analysis of Variance (ANOVA) and a chi-squared test.View in GitHub
The project consisted of providing insights for a fictitious Python-based ride-sharing app company, through an exploratory analysis on their data, creating visualizations to tell a compelling story.
The analysis is done utilizing Pandas, Scipy and Numpy, as well as Matplotlib for the visualizations, and showcases the relationship between the type of city and the number of drivers and riders, the percentage of total fares, riders and drivers by type of city, to help improve access to ride-sharing services and determine affordability for underserved neighborhoods.View in GitHub