I took 3 courses at Udemy over the last 4 years -
https://www.udemy.com/complete-guide-to-tensorflow-for-deep-learning-with-python
https://www.udemy.com/the-complete-sql-bootcamp
https://www.udemy.com/python-for-data-science-and-machine-learning-bootcamp
The course materials include data engineering, a variety of machine learning models from scikitlearn with corresponding theory from the book An introduction to statistical learning by James, Witten, Hastie and Tibishrani http://www-bcf.usc.edu/~gareth/ISL/ISLR%20Sixth%20Printing.pdf and building some neural networks with Tensorflow. Here I include two projects from the course. My ML and data engineering skills are now better than when I did the projects, but I have not updated the project notebooks. Overall I did about 10 projects per course, but I didn't feel like there was any value in including so many files. I am just trying to show some examples of the course level. My main goal in doing these courses was to learn python for machine learning and to learn tensorflow, which has some very complicated setup. The SQL course doesn't have any output that I can include in a github.
I use Anaconda with Python 3 and the following packages:
- tensorflow
- pandas
- matplotlib
- seaborn
- numpy
This analysis is of data that can be downloaded from Lendingclub.com . Those data are too big for github, so I have included an engineered version of the data here that removed some features, combined some features and removed NAs. It was beyond the scope of the project to do further feature engineering, but I might go back later to do that because the results are not the best.
This is the first project I did with tensorflow to predict 12 months in a timeseries.