Hey, I’m
Gabrielle
Beinars
Validation Scientist / Data Scientist
About Me
I currently work as a Process Validation scientist in the Process Development department at Nitto Denko Avecia, in Milford, MA. Nitto Denko Avecia is a contract manufacturing organization that specializes in oligonucleotides. Most of my daily activities involve technical writing, reviewing, and statistics. I will be receiving my master's in Data Science from Bellevue University in Fall 2021.
In my free time, I enjoy exercising, reading, and spending time outdoors with my partner and our dog, Bentley.
Work & Projects
The IMDB project is a multisource project where three separate datasets were cleaned and ultimately merged where the final dataset was stored as an SQL Database. Pulling and cleaning data from a flat file (csv), website and API were demonstrated. This project focuses on numerical and text cleaning, exploratory data analysis, and creating visuals.
Binary logistic regression is used to predict if a person will be diagnosed with heart disease based on a number of factors, including high blood pressure, high cholesterol and whether or not the person is a smoker. This project was completed in Rstudio.
Using python, exploratory data analysis was completed, various visuals created, and patterns/relationships were explored. Outliers were removed, and missing values handled appropriately. Probability mass function and cumulative mass functions were created to further explore the distributions and the percentile rank of certain values. Lastly, hypothesis testing was done, and a regression model created.
Using python, exploratory data analysis was completed, various visuals created, and patterns/relationships were explored. Outliers were removed, and missing values handled appropriately. Probability mass function and cumulative mass functions were created to further explore the distributions and the percentile rank of certain values. Lastly, hypothesis testing was done, and a regression model created.
A case study on volcanoes was conducted and the resulting data was used for this project. Various factors, including volcano type, location, eruption category and evidence, elevation, population, last eruption year and VEI (volcano explosivity index). Various techniques were explored, including exploratory data analysis, data cleaning, transformation, one hot encoding, logistic regression and results were evaluated using a confusion matrix.
This project focused on various statistical approaches to a dataset containing information on housing sales. Regression models were created and various metrics to evaluate and improve the model.
Many tasks were completed for this project, including confirming SSL certificate, reading HTML data from a URL, checking the status of a web request, decoding the reponse, parsing HTML, reading and parsing a JSON file using an API, and lastly, completing a data pull from the Twitter API.
Many tasks were completed for this project, including confirming SSL certificate, reading HTML data from a URL, checking the status of a web request, decoding the reponse, parsing HTML, reading and parsing a JSON file using an API, and lastly, completing a data pull from the Twitter API.
Data on forest fires was analyzed in order to predict the size class of a fire. Various models were explored, including random forest, KNN and gradient boosting, and accuracy results were evaluated. Predictions of fire size can also allow for prior action to be taken. Some of the questions that were answered are, whether or not wild fires are becoming more common, what is the most common cause of forest fires, and how accurately can the fire class size be predicted.