Hello World 👋 I'm Nicholas Wertz
My name is Nicholas Wertz, and I am a Data Scientist specializing in predictive models utilizing machine learning. I am a graduate of the Flatiron School's Data Science training program and I am experienced in Python, SQL, advanced mathematics, machine learning, data sorting, cleaning, analysis, and visualization.
- Speech Emotion Recognition - Capstone project for the Flatiron School
- In this project, I designed a Convolutional Neural Network model to identify emotion in speech from audio clips, utilizing keras
- Converted all audio to spectrograms for 5 emotional classes, using the librosa Python package
- Achieved 74% multi classification accuracy with minimal loss, applying keras’s ImageDataGenerator
- Obtained an F1 Score of 84% for the sleepiness emotional class, employing scikit-learn
- Twitter Sentiment Analysis of Apple and Google for AT&T - Collaborative project for the Flatiron School
- For this project, our team investigated public sentiment of products from twitter messages, using Python’s natural language toolkit
- Determined which company had more positive public sentiment, utilizing pandas and matplotlib on the provided data
- Produced an ensemble predictive model, combining Multinomial Naive Bayes and a Random Forest Classifier
- Forecasted the sentiment in tweets using this model, concluding that the client should stock more Apple products
- Water Well Functionality for the Tanzania Ministry of Water - Collaborative project for the Flatiron School
- For this supervised machine learning project, the team produced a predictive model for the functionality of water wells in Tanzania, primarily utilizing scikit-learn
- Inventoried and cleaned all data, removing any redundant features for modeling, using Python’s pandas library
- Constructed an ensemble model, integrating a decision tree classifier, logistic regression, and a KNN algorithm
- Reached an accuracy of 81% and a precision of 82% with the final optimized version of the model