sturdy-umbrella | project comes from attending a python meetup

sturdy-umbrella is a Jupyter Notebook library. sturdy-umbrella has no bugs, it has no vulnerabilities and it has low support. You can download it from GitHub.

the inspiration for this project comes from attending a python meetup where a presenter did a quick overview of data science using the ames, iowa data set. ames, ia housing data (xls). as a recent graduate of the galvanize immersive data science program i was exposed to many modeling techniques for which there wasn't enough time to delve into them deeper during the program. so i took this as an opportunity to work on scraping more data and merging it in. i also did a few rounds of experimentation with feature transformations and ensemble modeling. first, the metric used to evaluate error was rmsle, root mean squared log error. rmse is typical for models where the errors tend to be reasonably distributed. the log part is because of the log transformation that was performed on the price initially. this is done because the relative error is more useful than an absolute error. for example houses in this sample vary from 30k to over 500k. if the rmse (for convenience the 'average error') is 25k that is pretty good for a 500k house (~5%) but it's a horrible error for a 30k house (80%+ error). instead it's more useful to know something more akin to being 10% off on