Hello and welcome! I am happy and proud to share my Github profile and projects with you.
I am learning:
- SQL ๐ฉ๐ฝโ๐ป
- Data Structures and Algorithms ๐ป
- Japanese ๐ฏ๐ต
I am working on:
- Bot to source investment properties
- Automated investments membership platform
Inclusive Content Filtering For Social Media, 2021
Social media platforms like Facebook and Instagram have received a lot of public scrutiny regarding content filters in recent years. Image filters are regarded as too rigid and non-inclusive; Facebook as we all know has faced lawsuits regarding their inhibition on user freedom of expression. Following this outrage, I sought to build an image classifier for inclusive content filtering; a classifier capable of weeding our truly explicit images, without shaming nursing mothers or exhibiting racial bias. Truly a content filter for all.
SyriaTel Churn Predictor, 2021
Churn is one of the most important metrics a company can track and provides valuable insight into the customer experience and overall satisfaction with the company. SyriaTel's churn is 14x the churn of their top 5 competitors in the telecom industry. In my analysis of churned clients I deduced that a customer's "calls to customer service" and "day call charges" were key indicators of a customer's likelihood to "soon churn." Utilizing these features and several others, I built a machine learning model capable of predicting whether a client will "soon churn" with 83% recall.
Pneumonia Detector For Children Under 5, 2021
Pneumonia is a completely treatable illness yet still the leading cause of death for children under 5 worldwide. Time is of the essence once pneumonia has onset and in extreme cases a patient can die within hours. Radiologists however are overworked, overloaded, and burntout causing delays in diagnoses and increasing their room for error; the solution is found in AI assisted diagnoses. I built a convolutional neural network with 99% recall in my effort to ensure that no child should die from a treatable illness.
King County Housing Price Predictions, 2021
Zillow has faced controversy regarding their less than accurate Zestimate predictors; real estate professionals urge homeowners to not rely on Zestimates as indicators of home value and a group of home owners in Illinois filed a lawsuit against the company in 2017 for this very reason. In this project, my team and I built a regression model with Zillow Zestimates failures in mind. This predictive model was built using home data in King County, Washington from 2014-2015.
Microsoft Films Analysis, 2021
In this hypothetical scenario, Microsoft has decided to create a box office success and shake up the film industry with their very own Microsoft Films. My team and I were tasked with cleaning and analyzing data from thousands of films to determine what type of film Microsoft should create.
TD Ameritrade Automated Investment Bot, 2021
I built an investment bot to pull live stock prices every minute from the Yahoo Finance API during trading hours. If a stock's price crossed the 20 day moving average on a downward trend, a "sell" trigger went off for the bot to log into the user's TD Ameritrade account and execute a trade - selling all owned share's of that particular stock. When a stock's price crossed the 20 day moving average on an upwards trend, a trigger was set off for the bot to log into the user's TD Ameritrade account and execute a trade, utilizing all available funds in the user's account.