Python and Data Analysis Portfolio
Welcome to my portfolio of python and data analysis projects. Below are machine learning, data mining, and forecasting projects I have completed. One of my core values is life long learning and this portfolio is a record of some of the ways I have continued to learn post-grad.
Melbourne Housing Market Visualization – Tableau Public
Dashboard of housing data in Melbourne, VIC, Australia. The interactive visualization can be found on my Tableau Public profile.
Association Analysis Project – Online Retail II
View Code on GitHub: Association Analysis Project
View Jupyter Notebook via nbviewer: Association Analysis Project
This project explores association analysis, often called market basket analysis.
In learning about association analysis, I found that most resources typically took input as a list of products and records: (A, B, C),(B,C),(A), etc. In my experience this is not a common format for transaction data.
Another limitation of existing demonstrations is that they typically present small datasets, no more than a few hundred rows. Again, in my experience, datasets are rarely that small in business.
The goal of this project is to model an association analysis using large datasets (100,000+ records) and non-list formats of data input.
This project is written in python, using popular data science libraries pandas, numpy, and mlxtend. The algorithm used to find frequent itemsets is the fpgrowth algorithm.
Time Series Forecast Project – FRED Data – Advance Retail Sales
View Code on GitHub: Time Series Forecast Project
View Jupyter Notebook via nbviewer: Time Series Forecast Project
This project explores time series forecasting using the SARIMA model.
