Profile
Strengths include data-driven and analytical problem solving, quick learner, professional writing capabilities, excellent time management and prioritization, effective communication skills, self-motivated, and a team player.
Technical Proficiencies:
- Fluent in Python & R
- Database management with SQL
- Visualizations in Tableau and PowerBI
- Machine Learning & modelling using both supervised and unsupervised techniques.
- Specific experience coding and evaluating random forests, kNN, ensembles, SVM, linear regression, GLM, k-means, k-medians, NLP, Naive Bayes, meta-ensembles (model stacking), boosting (XGBoost), neural networks, CNNs, logistic regression, dimensionality reduction with PCA, linear discriminant analysis, and DBSCAN.
- Proficient with Keras and Tensorflow
- Familiar with Snowflake, Databricks, AWS, Docker
- Version control with Git and GitHub
- Graphing in R with ggplot and shiny, in Python with matplotlib and seaborn
Business Proficiencies:
- Report creation and presentation of conclusions to an audience or team (technical or non-technical audiences)
- Experienced in Microsoft Office, Adobe Illustrator & Photoshop, CRM tools
- Experienced in identifying and creating cost-effective strategies to improve online communication with customers
- Ability to collaborate with internal and external stakeholders to streamline communication and ease job completion
Work Experience
TransLink | November 2019 - Present | Analytics Professional in Training
As a member of the Research and Analytics department, I worked on projects that assist stakeholders in various departments throughout the company, including planning, marketing, finance, and business technology services.
- Created a machine learning model in Python that used DBSCAN, k-means, and LDA to segment over 2 million yearly Compass Card customers into 6 classes based on their weekly usage.
- Monitored and reported the stability of the classes, retention of customers, make-up of the classes by fare type, movement of customers over time, and growth/loss of the customer base by class.
- Became a source of ridership information and reporting throughout the COVID-19 pandemic using data from the segmentation model, along with other Compass Tap data to monitor and report the changes in ridership during COVID-19 by class of customer, time of day, fare payment method, and city of trip origin. This data has been reported to various stakeholders in planning, marketing, and the customer experience committee.
- Developed a PowerBI dashboard, SQL, and Python processes to automate COVID-19 related weekly and monthly overall ridership reports for various internal and external stakeholders, including Finance, Planning, CUTA, and UITP.
- Worked in a cross-functional team with Business Technology Services to test and review cloud-based data storage platforms, Snowflake and Azure, and their integration with Databricks and Tableau.
- Developed linear regression models to predict daily and weekly transit ridership using external data sources, with a mean average weekly accuracy of 99.3%. This data assists finance and planning with decision making while internal data reporting systems are offline.
Silvercreek Realty Group | 2015 - 2018 | Marketing Associate
Member of the Business Development Team that grew the company into a top-100 brokerage in the US (based on gross sales), increased the number of real estate agents at the company by over 60%, from 800 to 1300 agents, and obtained yearly production growth of over 20%.
- Responsible for tracking internal recruiting and retention statistics, identifying areas for improvement and presenting those findings to the BusDev and management teams.
- Tracked and reported monthly, quarterly, and yearly brokerage sales statistics.
- Mobile & electronic communications manager between the brokerage and the 1300 real estate agents.
- Designed, automated, and tracked performance metrics of internal and external marketing campaigns developed for either recruiting or retaining real estate agents.
- Assisted in website design and development for the internal and external brokerage websites.
- Core team member for the development of a 14-week agent training program and managed ongoing distribution of the program.
Education
University of British Columbia | Master’s in Data Science | 2019
Skills Acquired:
- Hypothesis development and testing using statistical analysis.
- Supervised & unsupervised machine learning for both regression and classification.
- Model performance estimation for regression: RMSE, MAE, R2; and classification: accuracy, precision, recall, AUROC
- Coursework using AWS, Docker, Tensorflow, Keras
- Version control with Git and GitHub
- Data Wrangling with SQL, Excel, APIs, web scraping (HTML & JSON)
- Data visualization in R with ggplot2 and Shiny, in Python with Matplotlib and Seaborn
Capstone Project:
- Partner company: Tetrad
- Team size: 4
- Project scope: Predict the success of a potential new franchise location for a restaurant chain.
- Language used: R
- Project description:
- Created a predictive analytics tool that leverages demographic, business, and traffic data in geographic zones to predict the success of opening a franchise location at a given location by predicting its potential profitability. This supervised learning model that can be used across multiple industries by any established B2C chain/franchise company.
- Feature selection: the algorithm scans through thousands of data variables and uses greedy selection to choose variables most predictive of success/failure for the company.
- Modelling: Meta-ensembling (model stacking) used to reduce error (+/- 15% of true value), prevent overfitting, and alleviate single-algorithm biases. Algorithm selects between linear regression and/or random forests regressor (based on lowest RMSE) for use in the final stacked model.
- Report generation: Produces automated report that estimates the quality of the predictions using statistical error measurements (RMSE, MAE, classification accuracy) and informs about the variables selected by the algorithm.
- Process:
- Extensive exploratory data analysis which included using graphing for feature visualization, correlation matrix, and linear/non-linear base models – in both R and Python.
- Explored linear and non-linear machine learning techniques including linear/logistic regression, random forests, SVM, boosted decision trees, and kNN – in both R and Python. Evaluated regression models using RMSE, R2, and MAE, evaluated classification models using accuracy, recall, and precision.
- Explored feature selection techniques based on Pearson correlation, regression weights, regularization, and greedy selection to pick approximately 150 strongest features from 2700 while maintaining interpretability.
- Created custom greedy selection algorithm that selects variables for a stacked (two-level) model without violating the validation set in the selection process.
- My role: Primary developer of our variable selection, stacked model, and report generator scripts.
- Note: Certain details of this project are protected by an NDA, but I am happy to answer questions about our project approach, steps taken in problem solving, and methods/algorithms used.
University of Idaho | B.S. Marketing | B.S. Human Resources Management | Minor in Advertising | 2015
Honours:
- Summa Cum Laude
- Dean’s List 2011-2015
- Recipient of the 2013 Alumni Award for Excellence
Curriculum:
- Coursework in marketing management, pricing, production, human resources hiring, interpersonal management, and teamwork.
- Integrated business curriculum included additional coursework in finance, accounting, project management, and information systems.
- Coursework in advertising storyboards, copywriting, journalism, and multi-platform campaign management.
Activities & Honours
Professional Track & Field Athlete | 2015 - Present
- 2017 Team Canada Athlete | 3,000m Steeplechase | 2017 IAAF World Championships - London, UK
- 2016 Olympic Trials Qualifier & Olympic Time Standard Qualifier | 3,000m Steeplechase
- Skills Developed:
- Superior time management
- Effective goal setting (short and long term), planning, and budgeting
- Ability to work as a team or as an individual to reach a goal
- Ability to perform under stress
OTFC Website Manager (Volunteer) | 2017 - 2018
- Responsible for the website upkeep of Oceanside Track and Field Club (Youth club based out of Parksville, BC)
NCAA Division I Cross Country and Track & Field | University of Idaho | 2010 - 2015
- 2014 and 2012 USTFCCCA Division I All- Academic Honoree
- Seven-time All-Conference Academic Honoree
- Four-time NCAA Preliminary Rounds Track & Field Championships Qualifier and 2015 Semi-Finals Qualifier
- Nine-time First Team All-Conference and Four-time Second Team All-Conference
Delta Zeta National Sorority | Active Member: 2011 - 2014 | Alumna: 2014 - Present
- Founding Mother of the Pi Kappa Chapter at University of Idaho
- Key committee member responsible for revision of chapter bylaws in 2014