Portfolio
Disaster Response Messages: Natural Language Processing
Completed
March 2019
Description
Implementation of a full data pipeline including ETL and supervised machine learning with natural language processing.
A web app was developed to allow users to enter their own messages to be classified.
GitHub Repo
Global Income Inequality
Completed
February 2019
Description
Development of a data dashboard using Flask for integration with html and Bootstrap for web development.
Data is obtained through an API query from World Bank Data and plots are developed with Plotly.
The app is hosted on Heroku.
Website
GitHub Repo
Customer Segmentation
Completed
November 2018
Description
Use of unsupervised learning techniques of Principal Component Analysis (PCA) and KMeans to identify customer segements and identify over- and under-represented segments.
Blog Post
GitHub Repo
Image Classification
Completed
November 2018
Description
Use of transfer learning to build a neural network capable of identifying 102 different flower types with over 85% accuracy.
PyTorch used to build the neural network; Python argparse library used to create executable training and prediction scripts.
Blog Post
GitHub Repo
Tell a Tableau Story: Homelessness in America
Completed
July 2018
Description
Interactive presentation of US homelessness data, including a brief look at causes and solutions.
Python (Pandas) used to clean; Tableau used to visualize the results.
Blog Post
GitHub Repo
Wrangle & Analyze Data: The World of Dogs
Completed
June 2018
Description
Data wrangling of the WeRateDogs tweets including gathering, wrangling and analysis.
Gathering included manual and programmatic downloads, and API access.
Python (Pandas, Numpy, Matplotlib, Seaborn, json, os, Requests, and Tweepy) used to clean, analyze and visualize the results.
Blog Post
GitHub Repo
Predicting Election Results
Completed
June 2018
Description
Explorations of contributions to the 2016 US Presidential Election candidates and the election results.
Includes univarite, bivariate, and multivariate EDA with multiple linear regression modelling and testing.
R, R Markdown and particularly the ggplot2 library used to visualize and analyze the results.
Blog Post
GitHub Repo
Analyzing A/B Test Results
Completed
April 2018
Description
Analysis of the conversion rate of a new website compared to an old, including the exploration of a potential interaction with user country.
Includes analysis using bootstrapping, traditional t-tesing and linear regression modeling.
Python (Numpy, Pandas, Matplotlib, and Scipy) used to visualize and analyze the results.
GitHub Repo
Female Labor Force Participation and Economic Strength
Completed
April 2018
Description
Examination of the relationship between country economic strength and female labour force and education participation.
Python (Numpy, Pandas, Matplotlib, and Scipy) used to visualize results.
Blog Post
Original Project
Updated Project
Exploring Bikeshare Data
Completed
February 2018
Description
Interactive user experience examining user statistics for bikeshare data in Chicago, New York and Washington.
Python used to code original project with code updated to use Pandas.
Blog Post
Code Walk-thru
Original Project
Updated Project
Exploring Weather Trends
Completed
February 2018
Description
Examination of temperature trends in six Canadian cities in comparison to global temperatures.
Excel used to chart trends and test temperature predictions.
Blog Post
GitHub Repo
Housing First Supports for Metro Vancouver
Completed
2016 - 2017
Description
Designed, applied for funding, and managed project examining support needs to implement Housing First in Metro Vancouver.
Website developed to support community resource identification and publication developed.
Website
Publication
Homelessness Resources
Completed
2014 - 2017
Description
As the senior staff member of the Greater Vancouver Shelter Strategy Society developed numerous resource to support the homlessness sector.
Projects included webinars, implementation guides, research reports.
Resources