Project Porfolio
Brandyn Ewanek
This page is meant to highlight selected personal data science projects completed by Brandyn. Highlighting a diverse set of skills in with predicting the stock with TensorFlow and the creation of a meta-model. A simulation of a team project between a data analyst and data scientist in the Spaceship Titanic Kaggle Predictions. Also highlighting the ability to extract business insights from customers and use it to create valuable marketing strategies for a supermarket.
Neural Network Stock Prediction - Meta-Model 3 Phase Development
2 of 3 Part Project factoring Economic, historical stock price, and Financial News in an effort to create a Meta-Model to predict the stock market. The long-term goal is to combine the 3 parts is one meta-model.
The goal of this study will be to determine the 30 most impactful economic features in predicting the daily close price of the Google stock, ticker symbol GOOG. A secondary point of study will be to define a deep learning model architecture Using the functional API in TensorFlow, for eventual inclusion in larger neural network
Day traders use various charting techniques to help in timing entry and exit into an investment position. Although difficult there is money to make in understanding the historical price movements, this must mean there is valuable information in historical stock movemets. With this our goal is to see if a network can refine this into a single time series feature for include in the main model.
SpaceShip Titanic Machine Learning Project
In this 2 part project for a Kaggle competition. We divid this isto a data analyst project and a machine learning engineer project.
​
This project has been turned into a follow along guided project for students to practice.
​
​
In the first part the data analyst we extract important insights to support that ML developer and the 2nd project will highlight how a model can be developed using the targeted insights extracted by a team mate.
In the 2nd In the second part of the Python Guided Machine Learning Project, the data scientist picks up where the data analyst left off. We use the data analyst's sights to guide the data scientist preprocessing strategy for machine learning.
​
We put many ML model through grid searches to find the best model but more to understand the hyperparameters we plot the entire grid search report in Seaborn's pairplot and give us a better sense which features really impact the models accuracy.
Extract Insights for Combination into Marketing Strategy
Designed to show students the power of staying focused on each task collecting insights and not until the end stepping back and being creative with those insights in interesting ways for a marketing campaign based on the data.
This project has been turned into a follow along guided project for students to practice.
​
​
tn this project I systematically go through each feature in this customer data set and extract insights staying focused on the specific task, and simply collecting insights. Then combining them for marketing strategies based on the data and designed to be profitable stratgies.