Brad-Swindell--Data-Science-Portfolio

Projects using Python, Pandas, SQL, R, Data Visualization Libraries, and Machine Learning Models

Click Project Title link below to see Jupyter Notebooks with detailed Python Code

Project 1: Binary Classification - Python code and Machine Learning Models for Prediction

Using input HR turnover data set from Kaggle web site: https://www.kaggle.com/giripujar/hr-analytics
Read in data via Python Pandas DataFrame, conducted initial review of fields and data analysis via Pivot Tables.
Converted String data fields to numeric using a look-up table for one and GetDummies command for the other.
Create Training and Testing datasets and run Logistic Regression, SVC, and Random Forest Machine Learning models for prediction.
Random Forest is best prediction model for this data with a score of 99.

Using Pitchfork Album Review SQL database from Kaggle web site: https://www.kaggle.com/nolanbconaway/pitchfork-data
The multiple-table SQLite database is read into a DataFrame with Inner Join statements in Python.
Exploratory Data Analysis is performed to analyze key variables and relationships.
Pivot Table created with the Genre of albums reviewed by Year.
DataFrame created from Pivot Table and format converted to “Regular” Dataframe for charting.
A Matplotlib line chart created showing album review Genres by Year.
DataFrame created with the top 4 record labels represented for visualization.
A Plotly Express “SunBurst” chart created to display Pitchfork’s top record labels and music genres for each one.

Class project created when student at NoVA Data School Data Science Boot Camp
Read in Data via R program code, conducted initial review of fields and data analysis via Pivot Tables.
Create Training and Testing datasets and run Linear Regression model for prediction.

Using Wildfire SQL database from Kaggle web site: https://www.kaggle.com/rtatman/188-million-us-wildfires
Exploratory Data Analysis is performed to analyze key variables in US Wildfires and relationships.
Assume Domain Knowledge to add a Category to the different types of Wildfire Causes.
Add Plotly Express “Tree Map” chart to visually show the categories and causes of wildfires.
Create Logistic Model to predict the Category of wildfires given other variables.