This Repository contains the five project portfolio completed during the Summer 2017 Metis bootcamp.
1. Strategic Volunteer Placement
2. Bridging the Achievement Gap
Background: An analytics firm is approached by a non-profit advocacy organization for women to be greater represented in the tech industry. Every year, they hold a gala to bring awareness to their cause. The organization wants to maximize the amount attendees that attend the gala this year and request the help of the firm.
This Project uses data from subway turnstiles provided by the Metropolitan Transportation Authority of New York.
The goal of this project was to use the turnstile data to recommend around which subway stations volunteers should be distributed to maximize the number of attendees to this years gala.
Note: This project was a one-week hands-on-experience exercise in becoming proficient with the Python pandas library as this would be a basic tool for introductory data exploration going forward in the bootcamp.
I'm proud of the webscraping script I created in this notebook. Anyone looking at this repo, should definitely take a look at this project.
As a former high school science teacher, student achievement was very important to me. A lot of academic literature has been published on something known as the "Achievement Gap", a noticeable difference in test scores between students of two demographic groups.
Though it is impossible to get data on individual students due to student educational privacy laws, a lot of aggregated data is publicly available. Using data scraped from the Illinois Report Card website, I attempt to predict the Hispanic-White Achievement Gap of a school using a multiple linear regression model.
The goal of this project was to classify the satisfaction of citizens with their lives based on their answers to certain survey questions. Raw data was collected from the 2014 PEW Global Survey results.
Given the explosion of "fake news" available in our society today and its toxic effect on political discourse, I attempt to use unsupervised learning to cluster the fake news published around the time of the 2016 Presidential election into distinct categories.
Despite the current political climate (as of Feb 2018), I still carry (non data-driven) hope that this phase of disgraceful political rhetoric shall come to pass.
A few days before I began this project, my brother had gotten into a pretty severe car accident which totaled his car. I wanted to find out the factors related to the severity of collisions to propose data-driven recommendations to state agencies for transportation safety improvements.