Marley Buchman

Consultant providing data science services and applied economic research for public and private sector clients. Managing projects, developing methodologies, and delivering quantitative analyses improving client’s decision making and processes.


Experience

Economist // ECONorthwest // Portland, OR // June 2017 to Present

Provided data science consulting services and applied economic research for public and private sector clients.

  • Developed a Random Forest model using scikit-learn to predict multifamily rent in Portland, OR in order to calibrate land use policy

  • Developed three R packages that saved analysts’ ~100 hours of time by automating an annual report, and making it easier to work with CoStar real estate data

  • Created a bespoke spatial interpolation algorithm in Python for custom land use modeling software (Plan4Dev)

  • Estimated the effects of pollution on property values with a spatial econometric model in R that was used as evidence in a litigation case

  • Built a Shiny analytics dashboard for a commercial real estate developer that was used to inform real estate acquisition decisions

  • Was the primary quantitative researcher on award-winning research on national housing underproduction that was featured in many journalistic outlets such as the Wall Street Journal and CityLab

Data Science Consultant // Trucksmarter // Remote // June 2021 to Present

Developed and deployed proof of concept machine learning model to predict price of loads for a trucking logistics startup.

  • Developed a XGBoost model in Python to predict the price of a trucking load based on historic trucking data

  • Created data pipeline to process raw JSON data and write to Postgres database

  • Created an API to serve the model using FastAPI and dockerized the API using Docker

  • Deployed the dockerized API on AWS (EC2) using AWS ECR to host the container

Product Analyst Consultant // Trellis Technologies // Remote // October 2020

Developed model to optimize expected ad revenue per search for a FinTech startup creating a car insurance marketplace app.

  • Developed a linear regression model to estimate CTR with a RMSE of 1.6%

  • Back tested the model and estimated that the model would increase expected ad revenue by 0.41% or $113/day

  • Performed exploratory data analysis to identify trends and insights from existing CTR data in Looker

Senior Economic Associate // REMI // Amherst, MA // November 2015 to May 2017

Provided R&D, training, and consulting for an economic modeling software company.

  • Advised and consulted on studies such as the economic impact analysis of high-speed rail project in California, Measure M: LA County’s T.I.P., Alaska B.R.A.C. proposal, Washington tax incentive programs, etc.

  • Assisted in the research and development of dynamic economic simulation software

  • Modeled economic and fiscal impacts of changes in transportation, taxation, energy, and environmental policies for clients

  • Presented talks/seminars/webinars on Dynamic Economic and Demographic Forecasting, The Economic Impacts of Lifting the Crude Oil Export Ban, Demographics as Destiny, etc.

  • Advocated for the use of economic models and data driven analysis to state legislators, foreign govt. officials, state and local govt. directors, academics, business groups, etc.

  • Business development, software training and technical support

Education

2013-2015

Msc, Economics; Uppsala University, Uppsala, Sweden

Thesis title: The Bakken Transformation: The Evolution of Criminal Behavior in the Wake of a Resource Boom

Analyzed how criminal behavior changed in response to a large, exogenous positive shock to the regional labor market

2009-2012

BSc, Economics, Political Science; University of Oregon, Eugene, OR

Talks, Presentations, & Open Source Work

Organizer // Portland R User Group // Portland, OR // March, 2019 to Present

Organized monthly meetups for R programmers and Data Scientists in Portland Oregon.

HUD-MFI-API:

Developed and deployed an open-source API using Python, Postgres, and Flask to access historical income limits from the U.S. Department of Housing and Urban Development website for affordable housing analysis.

Skills

R, Python, Bash, SQL, Postgres, GIS, pandas, numpy, sklearn, Flask, FastAPI, Shiny, tidyverse, AWS Lambda, AWS RDS, AWS EC2, docker, Currently learning Airflow

michaelmbuchman@gmail.com

Portland, OR