Brent McRoberts

  1. Home
  2. Texas
  3. Texas A&M University
  4. Brent McRoberts
Brent McRoberts

Brent McRoberts

  • Courses0
  • Reviews0
Rate This Professor

Social Profile:

Join to see

Biography

Texas A&M University College Station - Geography

Advanced Statistical Modeler at Direct Energy
Environmental Services
Brent
McRoberts, PhD
Richmond, Texas
Professional, dynamic, results-oriented data scientist with 15 years of experience combining extensive technical, statistical and interpretive skills to deliver insights and implement action-oriented solutions to complex problems. Experienced at creating data regression models, using predictive data modeling, analyzing data mining algorithms, data visualization and communicating complex information in an understandable way.

Experience

  • Texas A&M University

    Research Scientist

    Created GIS maps still used for risk assessment and data visualization of drought indicators for the state of Texas.

    • Headed Texas drought calls, and advised on drought management strategy for the state of Texas, including the disbursement of billions of federal aid for the state of Texas
    • Improved accuracy in distributing billions of federal aid for the state of Texas by using big data, GIS, and statistical analysis to improve a drought monitoring product
    • Revitalized the Texas State Climatologist website (http://climatexas.tamu.edu/), redesigning data visualization techniques to include statistical and informational graphics

  • Texas A&M University

    Postdoctoral Fellow

    Developed risk management models and utilized technical writing ability to secure multiple grants and write multiple publications. Maintained several high-resolution big data sets and mentoring numerous students doing undergraduate research.

    • Awarded $510,000 NOAA grant to develop software platform to integrate improved data into a NASA modeling framework, resulting in a promotion to Research Assistant Professor
    • Collaborated with another University to build the National Soil Moisture Network database (http://nationalsoilmoisture.com/) with several thousand weather stations and several gridded, interpolated products.
    • Mentored and motivated undergraduate students in research that influenced several students to pursue graduate degrees

  • Texas A&M University

    Research Assistant Professor

    Utilized cloud computing, machine learning, data analysis for modeling predictive risk assessment. Responsible for teaching upper level statistics, data analysis, and GIS courses.

    • Reduced biases in land surface model estimates of soil moisture by 700% through assimilation of high-resolution satellite data
    • Consulted as software developer for a power company responsible for 14,000,000 people by building a machine learning (Random Forest) algorithm to operationally predict severe weather-related power outages
    • Lowered modeling error for hurricane related power outages by 20% by developing a predictive power grid outage model
    • Awarded $100,000 National Science Foundation grant to develop software algorithms to assess rainfall associated with Hurricane Harvey for improving risk assessment
    • Collaborated with scientists to determine that a Texas evapotranspiration network would provide $25+ million worth of benefits for Texas agriculture
    • Partnered with USDA/NRCS to develop improved annual planting zones in Texas

  • Direct Energy

    Advanced Statistical Modeler

    Developed six complex machine learning algorithms to provide actionable marketing insights for Direct Energy’s acquisition and retention campaigns.

    • Created a machine learning, look-alike algorithm to determine which of the 50,000,000+ customers in the energy footprint should be targeted for direct mail advertising
    • Rebuilt a Return on Marketing Investment (ROMI) model to attribute energy sales to the $10,000,000+ spent on different marketing channels
    • Developed a new propensity model to properly identify which of the 100,000+ customers in Alberta are likely to churn, triggering target retention efforts
    • Identified Wal Mart locations in Ohio and BJ’s Wholesale Clubs likely to have success selling retail energy using a machine learning model

Education

  • Texas A&M University

    Master of Science - MS

    Atmospheric Sciences and Meteorology
    • Thesis: Drought over the past century in Texas and New Mexico: Reducing inhomogeneities in long-term climate records via statistical methods to study drought • Developed a new climate division precipitation dataset (FNEP) to homogenize the precipitation record in Texas and New Mexico. This dataset has been expanded to cover all of the continental United States.

  • Texas A&M University

    Doctor of Philosophy (Ph.D.)

    Atmospheric Sciences and Meteorology
    Dissertation: Minimizing Biases in Radar Precipitation • Estimate • Wrote an extensive suite of software algorithms that ingests radar precipitation data (4 km resolution across the CONUS) and detects biases through spatial pattern recognition and ground-truth data • Outstanding Graduate Student Teaching Assistant Award

  • Texas A&M University

    Research Scientist


    Created GIS maps still used for risk assessment and data visualization of drought indicators for the state of Texas. • Headed Texas drought calls, and advised on drought management strategy for the state of Texas, including the disbursement of billions of federal aid for the state of Texas • Improved accuracy in distributing billions of federal aid for the state of Texas by using big data, GIS, and statistical analysis to improve a drought monitoring product • Revitalized the Texas State Climatologist website (http://climatexas.tamu.edu/), redesigning data visualization techniques to include statistical and informational graphics

  • Texas A&M University

    Postdoctoral Fellow


    Developed risk management models and utilized technical writing ability to secure multiple grants and write multiple publications. Maintained several high-resolution big data sets and mentoring numerous students doing undergraduate research. • Awarded $510,000 NOAA grant to develop software platform to integrate improved data into a NASA modeling framework, resulting in a promotion to Research Assistant Professor • Collaborated with another University to build the National Soil Moisture Network database (http://nationalsoilmoisture.com/) with several thousand weather stations and several gridded, interpolated products. • Mentored and motivated undergraduate students in research that influenced several students to pursue graduate degrees

  • Texas A&M University

    Research Assistant Professor


    Utilized cloud computing, machine learning, data analysis for modeling predictive risk assessment. Responsible for teaching upper level statistics, data analysis, and GIS courses. • Reduced biases in land surface model estimates of soil moisture by 700% through assimilation of high-resolution satellite data • Consulted as software developer for a power company responsible for 14,000,000 people by building a machine learning (Random Forest) algorithm to operationally predict severe weather-related power outages • Lowered modeling error for hurricane related power outages by 20% by developing a predictive power grid outage model • Awarded $100,000 National Science Foundation grant to develop software algorithms to assess rainfall associated with Hurricane Harvey for improving risk assessment • Collaborated with scientists to determine that a Texas evapotranspiration network would provide $25+ million worth of benefits for Texas agriculture • Partnered with USDA/NRCS to develop improved annual planting zones in Texas

  • Purdue University

    Bachelor of Science - BS

    Atmospheric Sciences and Meteorology
    • Undergraduate Thesis: Relationship between aircraft-related turbulence and synoptic meteorological data • Coordinated observed turbulence on traveling aircraft with meteorological conditions to find significant statistical correlations