Drew Hogg, PhD

Data Science


Machine Learning
Computer Vision
Large-Scale Simulations
Natural Language Processing


Ph.D. Astronomy, University of Maryland
M.S. Astronomy, University of Maryland
B.A. Physics, University of Colorado
B.A. Astronomy, University of Colorado



Recent Projects

Industrial Drone Startup Company
Structural Fault Detection
Developed a system using a mix of traditional computer vision techniques and deep learning techniques to detect faults such as cracks in industrial installations. The data was from videos captured from large drones.

Social Media Analysis Company
Geographical Inference from Social Media Posts
Analyzed both the text and imagery of social media posts to determine if they give hints at where the person posting is physically located. A deep learning computer vision approach was used to detect clearly identifiable landmarks in the backgrounds of photos. Also, with a pre-trained hybrid bi-directional LSTM and conditional random field model, named entities were identified and extracted from the foreign language text documents.

Technical Expertise

  • Spark, SQL, Hadoop
  • Python data stack: scikit-learn, pandas, numpy, Jupyter, others
  • Scientific analysis, statistical analysis, numerical modeling
  • Regression analysis, time-series analysis, Bayesian statistics
  • Deep learning on Tensorflow and pytorch