Data Science Lab

Teaching Photo 

An introduction to Machine learning and Data science for engineers. (Updated logistic information on UT Canvas)


Constantine Caramanis
Alex Dimakis


Machine Learning and data science are transforming engineering, healthcare and scientific discovery. In this class we are going to discuss how to use data to build models for prediction and inference. We put a special emphasis on engineering applications, signal prediction and modeling.
Predictive modeling. Regression and Classification. Data cleaning and preprocessing. Feature engineering. Unsupervised methods. Data clustering. Model selection and feature selection. Entropy and Information theory. Deep learning models. Multilayer Perceptrons and Convolutional Neural networks. Machine learning for signals and timeĀ­series data. Spectral features and Fourier Transforms

This class has a significant applied component and involves working with real data, performing modeling and prediction. Python, Numpy, Pandas, Scikit and Tensorflow required, please use the Anaconda distribution. Additional tools will be discussed when introduced.

Data Exploration in Pandas

In the first lab we warmup in Python and Pandas by exploring a mysterious dataset that is given to us without information about what the features are. We perform basic data exploration and visualization in Pandas.

See the ipython notebook

Model preparation and Tuning for Kaggle

Model Tuning Notebook

This notebook uses the data from the Advanced regression kaggle competion Advanced Regression techniques

Additional content

Some Final Projects from Spring 2017:

Most updated information on UT Canvas site.

Funding from Intel, Amazon, Google and Microsoft supports the development of this course and is gratefully acknowledged. Student cloud credits were generously donated by NVIDIA and Microsoft Azure.