Barrett Duna’s Resume
Data Scientist Professional
Barrett Duna is a professional data scientist with an 8 year history of exceptional data analytics performance. Starting at Live Nation Entertainment as a data analyst analyzing a database of tens of millions of ticket transactions for the purpose of optimally pricing tickets and forecasting ticket sales. Moving on to his 7 years of experience at Duna Analytics working on various data visualization, machine learning and AI projects. All while spending time in marketing and owning a digital marketing agency for three years serving various Bay Area clients.
Barrett Duna's education includes a B.S. in Mathematics/Economics with Ph.D. level coursework in Applied Sampling, AI/Machine Learning and Econometrics.
Currently, Barrett is a newly admitted student at George Mason University starting in the Fall of 2020 studying an M.S. in Data Analytics Engineering in Volgenau School of Engineering. The program specializes in big data and is recognized as one of the nation's top 25 big data programs. Coursework will cover Big Data, Database Management, AI, Machine Learning Intro and Advanced, Data Mining, Predictive Analytics and the core of the program.
Machine Learning / AI Algorithms:
Regression Analysis, KNN, K-Means Clustering, Logistic Regression, Deep Learning Neural Networks, Time Series Analysis and Forecasting, Adaptive Linear Neurons (Adaline), Convolutional Neural Networks (CNN), Generalized Linear Models (GLM), Principal Component Analysis (PCA), Random Forests, Gradient Boosting Machines (GBM), XGBoost, Support Vector Machines (SVM), Econometric Techniques, Gradient Descent, Mini-Batch Gradient Descent, Stochastic Gradient Descent (SGD), Applied Sampling Techniques, Optimal Pricing and others.
Technical Tools, Skills and Languages:
Python (Fluent), R (Proficient), SQL (Proficient), Data Visualization, Microsoft Office Suite including Excel, Tableau and others.
George Mason University
M.S. Data Analytics Engineering
Newly admitted student. Starting fall 2020. Program focuses on the application of data analytics on big data. Coursework spanning big data technologies to Machine Learning/AI will be taken.
B.S. Degree in Mathematics/Economics. Featured the core of both the
Mathematics degree and the Economics degree. I focused my elective
coursework in Mathematics, Statistics, Probability Theory and Data Analytics related courses.
Relevant Coursework: Calculus Series, Lower Division Linear Algebra,
Elementary Statistics, Introduction to Programming C++, Matlab for
Engineers, Discrete Mathematics, Differential Equations, Upper
Division Linear Algebra, Introduction to Probability Theory, Real
Analysis, Introduction to Mathematical Statistics, Introduction to Econometrics, Mathematical
Optimization, Advanced Real Analysis, Introduction to
Computation and Optimization for Statistics, Abstract Algebra for
Three Ph.D. level courses: Applied Sampling, Machine
Learning/AI and Econometrics.
Duna Analytics / Data Scientist
January 2020 - Present - Santa Rosa, CA
Utilized Python and R to perform data analysis, visualization and build models using AI and machine learning algorithms. Projects included analyzing customer demographics and producing visualizations, real estate housing price analysis, optimal pricing analysis based on economic transactions, etc.
One project involved using a mixed multinomial logit model to model B2B service choice decisions and produce a report on the determinants of what
causes a business to choose one service alternative over another based on a multinomial dataset collected in a marketing department on business service choice factors.
Great Life Digital / Founder and Chief Digital Strategist
April 2017 - January 2017 - Santa Rosa, CA
Handled all aspects of running a company including administrative, legal, tax, marketing and sales tasks. Primary responsibility was sales but I used my analytics skills to aid clients.
For example, I built a marketing A/B test (testing two versions of web pages, landing pages, app interfaces, etc.) script in R which tests the statistical significance of an A/B test. This isn’t a trivial task. It actually involves running simulations on the A/B test possible outcomes to calculate a p-value.
Basically, larger sample sizes (traffic) for both versions A and B and larger differences in conversion rates lead to strong statistical significance while the opposite scenario fails to be statistically significant. See the explanation and code here:
Duna Analytics / Data Scientist
July 2015 - April 2017 - Santa Rosa, CA
Programmed in R and Python to analyze various data analytics projects and build appropriate predictive models.
Projects included data visualization and analysis for many different business applications. Academic scientific research analysis including effectiveness of different treatment levels in soils and plant growth. Many projects utilized regression analysis to predict continuous variables. There were projects forecasting time series.
In one case, the time series was factory production where the client wanted to be able to forecast production, but there was missing production data that the client wanted imputed.
Ygrene Energy Fund / Marketing Intern
December 2014 - July 2015 , Santa Rosa, CA
Performed general marketing duties. Applied analytics every chance I got.
Helped understand and optimize conversion rates for direct mail, email, mass fax and SMS. Analyzed consumer demographic data to help us better understand our customers. Segmented customers based on behavior and demographics.
Major focus was on determining optimal addresses to send out direct mail pieces to based on household characteristics. See:
Duna Analytics / Data Scientist
June 2013 - December 2014 - Santa Rosa, CA
Programmed in R and Python various machine learning, AI and data analytics solutions to complex business problems.
ANOVA analysis, deep learning neural nets for computer vision, deep learning neural nets for time series analysis, etc. Also, did data visualization to communicate to decision makers complex data analytics results.
Live Nation Entertainment / Data Analyst Internship
July 2012 - June 2013 , Hollywood, CA
At Live Nation Entertainment, was part of the live analytics department. The department was responsible for providing analytics to the major sports teams and
concert venues. Analyzed a database with tens of millions of ticket transactions.
Responsible for finding patterns in the data, visualizing those patterns and communicating those patterns to management and our clients.
Analysis included: effects of seating row (front versus back) on resale pricing, visualizations of resale price distributions by seating sections, identifying ticket brokers (individuals that buy up bulk tickets the second they go on sale and resale for a profit) and many other research investigations.
Also, optimally priced tickets to maximize revenue and profits. Used the knowledge from the Research Assistant position at Anderson School of Management (UCLA’s MBA program) to help determine profit maximizing prices looking at elasticity of demand and our databases of resale ticket transaction
One example of pricing adjusted was the 2013 New York Yankees bleacher seats. They were being sold for $15 in 2012 to cater to low income individuals in New York who couldn’t afford more expensive seating. Identified that bleacher tickets were being immediately purchased by brokers based on an extremely high volume of transactions within hours of the tickets going on sale. Then looked at our resale market transaction data and found that tickets were being resold for $45 on average for an average $30 profit for the brokers. A recommendation was made to increase price to the true market value (average resale price) of $45. This increased the Yankee’s revenue and eliminated broker profits.
One project involved reverse engineering the SF Giants dynamic ticket pricing by applying regression analysis to their prices using dynamic factors such as day of week, team performance, player pitching, etc.
Anderson School of Management UCLA / Research Assistant
December 2010 - June 2012 Los Angeles, CA
Developed a computational simulation of the primary and secondary (resale) ticket market in Python. The task involved converting an academic paper
with a 30-page advanced Economic model using Mathematics and Probability Theory into a working computer simulation. The goal was to study the optimal pricing of tickets, ticket market dynamics, eliminating broker profits, primary and secondary market interactions and general study
of the markets.
Coded the project in Python and it was a massive year and a half long project. The simulation involved various economic agents (consumers, brokers, the organization selling the tickets, etc.) and each economic agent
had a set of parameters that determined how they behaved in the simulation. There were also general model parameters e.g. number of consumers and brokers, agent behavior probability distribution parameters, etc.
Created a simulation where these agents interacted and the outcome of the simulation was based on the parameter values. The idea was to run the simulation tens of thousands of times changing parameters and
pricing based on simulation outcomes until a Nash Equilibrium is reached (each agent is responding to all other agents in his or her best interest) and revenue is maximized.