terrajriley / mini_portfolio_mt

This repository is specifically for showing relevant experience to those at the Mason Tillman firm.

Mini_portfolio_MT

This repository is specifically for showing relevant experience to those at the Mason Tillman firm.

De-duplication of data (model: DemandTools)

In my Capstone, I set up code such that it would identify which videos boths channels had talked about, matched them and then went out to get the comments for those videos. A summary of this process can be found here.

Also, in Toxic Watch Project I’ve set it up to avoid redundancies within the current Scrapings.

Automate classifying contracts by the type of purchase using North American Industry Classification System (NAICS) code. This has previously been done by querying text fields for keywords

I have created a number of classification models and am curious to learn more about the NAICS.

Twitter Neural Network

Predicting if Loans would be given or not

A crawler application to search various non-standard websites for ethnicity, gender, and maintain the sources of such data (model: import.io)

I’ve done web scraping of a number of sites including the Overwatch Forums, wikipedia pages, Seeking Alpha and Amazon with a number of tools including the requests, beautiful soup and selenium python libraries as well as dexi.io. The majority of my webscraping work can be found here.

Automate generation of a data quality report detailing consistency, completeness, accuracy, and missing variables

I always address missing variables when working with data and believe that generating reports on the consistency, completeness and accuracy of a dataset would be a fun and straight forward task.

Automate matching field headers from various sources to the field headers in company database

Before deciding to shelve a specific project I was developing a system (for that project) to automatically sort various different ways to say state the same value. I was doing this by creating a dictionary that would use the keys as the various ways that people have stated something and the values as the way that we wanted it to be stated. I would add any new values that I came across as the I would then use this dictionary as a reference

Automated and dynamic reports/dashboards. A summary report listing the most reliable “winner” variables (for each contractor or contract) but maintain all other variables in a master database.

An example of my EDA work can be found here I have also used AWS, primarily for my Toxic Watch project, and used code to update a dataframe table after every run of a Machine learning model to keep an eye on my progress as I've tuned the model with different featuresand hyper parameters.

This is a direct link to my Toxic Watch project which is referenced a few times in this document. In it I set up an automatic scraper of the Overwatch Forums with the intension to compair the toxic behaviour of players to the steps that the Overwatch Team has taken to address and eliminate said toxicity.