I worked on the Deep Analytics team identifying and acting on opportunities to apply machine learning on different areas of the company to increase revenue and make processes more efficient.
Initiatives examples :
- NPS modeling : understand how the network variables affect the NPS score of the company and use them to predict the detractors.
- Click Attribution : Comprehend the customer journey step-by-step using a
Markov Chain approach
-Court Decisions : Predict the probability of a lost cause as well as the amount of the sentence involved in the legal case so Nextel can attempt to sign agreements when it makes sense
- Complains : Using variables linked to the use profile of the customers, identify cases of high complain probability and take action preventively
- Network degradation study : Comprehend the concepts of KQIs such as RTT,
Lost Packages and Retransmission rate. The study identified spots where these indicators needed to improve, ploting all of them in maps KML and using them as features to explain events like churn and complains using
XGboost and Shapley Value
At Nextel out team works consistently with tools like :
- Python/R as statistical tools using libraries like pandas, matplotlib, scikit, dplyr, ggplot and caret
- XGBoost, Random Forest, Shapley Value, Logistic Regression, Time Series as algorithms for statistical modeling
- SQL Impala/Hive in a data lake environment to data extraction
Movile
Data Scientist
julho de 2016 - fevereiro de 2018 1 ano 8 meses
Campinas
I worked on the Machine Learning Team developing statistical models and data analysis to support decisions of several internal clients on Movile
I used tools like R, Python, Microsoft Excel and SQL to extract as much value as possible from data and improve results.
A few examples of Data Science projects that I worked are : Client
Segmentation RFM analysis, Churn probability Survival Analysis and ETA for delivery services Linear regression, XGboost, Deep Learning, KNN .
Telefônica Brasil
Senior Big Data Analyst
março de 2016 - julho de 2016 5 meses
São Paulo e Região, Brasil
Development of classification models Logistic regression, Decision tree and
Random Forest oriented for churn.
Manipulation of big data sets using SAS, Teradata SQL and R.
TIM Brasil
Geomarketing consultant
julho de 2015 - março de 2016 9 meses
São Paulo e Região, Brasil
Development of models focused on understand behaviour based on
geolocation and sociodemographic variables. The main focus was in cluster analysis to predict buying propensity in home broadband.
Softwares used : Sas Miner, R and Microsoft Excel
Value Partners
Intern
setembro de 2013 - dezembro de 2013 4 meses
São Paulo Area, Brazil
I worked as an intern providing support to the consultants in projects in different segments such as services, oil and gas and chemicals.
Itaú Unibanco