Data Science Portfolio

• Artifical Neural Network with Backpropogation
  
• Baby Name Time Series
  
• Convoluted Neural Network for Image Recognition
  
• Movie Recommender Using Cosine Similarity
• Predicting Cryptocurrency Values
• Neural Network for Image Classification with MNIST Data
• Movie Recommender Using Non-Matrix Factorization
• Pillow Tutorial
• Predicting Baby Name Popularity Over Time
• Predicting Survival on the Titanic
• Lyrics in the Style of

Artifical Neural Network with Backpropogation

In this file and this file, I construct a neural network from scratch with backpropogation.

Baby Name Time Series

In this file, I look at the time series data of baby name popularity. Data looks at baby name popularity over time and the number of distinct names over time as categorized by gender. I then use linear regression and polynomial regression to predict the popularity of specific names over time.

This project uses baby name data supplied by the U.S. Social Security Administration that can be found here.

Convoluted Neural Network for Image Recognition

In this file, I build a neural network to classify elements within images.

Movie Recommender Using Cosine Similarity

In this file, I build a movie recommender that utilizes cosine similarity. Cosine similarity gives a measure of similarity between users based on ratings they have provided for other films.

Predicting Cryptocurrency Values

In this file, I extract cryptocurrency values using the Coingecko API and look at time series data. I then predict future values using a number of machine learning methods, including linear regression, polynomial regression with Ridge and Lasso regularization, and Prophet.

Neural Network for Image Classification with MNIST Data

In this file, I build a neural network to classify the MNIST data set of handwritten figures.

Movie Recommender Using Non-Matrix Factorization

In this file, I build a movie recommender that utilizes non-matrix factorization.

Pillow Tutorial

In this file, I give a tutorial for the uses of the Python Library Pillow and build a GIF.

Predicting Survival on the Titanic

In this file, I use information given for the passengers of the Titanic to analyze the best predictors of survival. This submission received of score of 77.99% on Kaggle.

Lyrics in the Style of

In this file, I import artist lyrics from lyrics.com then vectorizer them and use bag of words to determine lexicons and styles of specific musical artists. Lyrics can them be inputed into the program and a 'most probably artist' returned based on this assessment. Scattertext is used to display results.