Michael Markin | Data Scientist

Hello, I'm Michael Markin.

I do data
And I strive within the intersection of finance and computer science.

About

To make a meaningful impact in the workplace (and the world), it takes more than just being a capable programmer. It also requires someone who will ask creative questions after fully understanding the workflows and data around them.

I am a passionate Data Scientist with nearly 3 years of working experience.

Recently, my focus has been on unsupervised Natural Language Processing (NLP).

I have graduated with a B.S. in Finance & Business Analytics with Honors and a minor in Computer Science.

I am currently developing a neural network error correction model to autocorrect writing mistakes.

I actively practice technical analysis stock trading.

Fun fact: I first got into coding to create a trading algorithm in Python.

I am self-motivated, always learning, and driven to excel.

Experience

NewLane Finance

Data Scientist (Credit Risk & Lead-Scoring)

I collaborated with the VP of underwriting to analyze and report on credit risk trends. I also worked with the lead data scientist to advance the company's lead-scoring model using NLP.

September 2020 - January 2022 | Philadelphia, PA

Glenmede

Data Scientist Co-op (Fixed Income)

I worked alongside traders and portfolio managers to automate workflows and build custom dashboards from portfolio data.

September 2019 - March 2020 | Philadelphia, PA

Schnader Harrison

Data Analyst Co-op (Marketing)

I joined the marketing team to provide insights on website traffic. I also developed a Python script to webscrape relevant business proposals from county websites.

September 2018 - March 2019 | Philadelphia, PA

Projects

Click on a card to expand it.

                
NJ Beach Tracker
                  
                  A website that helps people find the best New Jersey beach on any given day.
                
Accomplishments
                  The site provides a live 5-day report of weather forecast information for 28
                  of New Jersey's most popular beaches. Driving times to each beach can also
                  be compared once an address is entered.
                
Sentiment Predictor
                  A Dockerized Flask app that predicts sentiment on user-inputted text.
                
Accomplishments
                  This project deploys a simple sentiment prediction interface using Flask and Docker.
                  Once a user enters a valid text input, sentiment is automatically
                  predicted and displayed as POSITIVE or NEGATIVE along with the model's
                  confidence score for that prediction.
                
                  Custom NER Model
                  Auto-detecting home improvement services within unstructured text.
                
Accomplishments
                Building a custom Named Entity Recognition (NER) from scratch using the latest version of spaCy v3.
                After annotating and training on a mini dataset, the model is able to identify
                various types of home improvement and contractor services within unstructured text.
                
Python-to-Excel Techniques
                  A Python walkthrough demonstrating methods of writing to existing Excel files.
                
Accomplishments
                  I show how to use the Openpyxl library to write data to existing Excel workbooks,
                  automatically modify cell styling/formatting, as well as to
                  insert Python graphs as images to Excel.
                
Urban Dictionary Bot
                  Automatically search slang and upvote on UrbanDictionary.com using a ChromeDriver.
                
Accomplishments
                  The script allows a user to input a slang term,
                  which will then be automatically searched and upvoted on UrbanDictionary.com using
                  Selenium and a ChromeDriver.
                
E-commerce Reviews Classifier
                  
                  Visual analysis and model training on a dataset of clothing reviews.
                
Accomplishments
                  A supervised text classification model is built using various
                  text preprocessing methods, TF-IDF, and an ensemble of Scikit-learn models.

Articles

I am actively writing on Medium.
Check out my recently published work.

Skills

Languages

• Python
• JavaScript
• HTML/CSS
• R
• Bash
• Bloomberg BQL
• SQL

Web Development

• Node.js
• Express
• Flask
• Axios
• Bootstrap
• DataTables

Tools & Technologies

• Excel & Analysis ToolPak
• Tableau
• Jupyter Notebook
• Git/GitHub
• Microsoft SQL Server
• Docker
• Heroku
• Streamlit

Python is my language of choice.
I am experienced with data processing, web scraping, and machine learning projects.

Below are my most frequently used libraries.

Data Processing: Pandas, NumPy, Regex, Matplotlib, Seaborn, Openpyxl
Web Scraping: BeautifulSoup4, Selenium, Requests
Machine Learning: Scikit-learn, NLTK, Transformers, SpaCy, XGBoost, Gensim, Tensorflow, Keras

Education

Drexel University

Philadelphia, PA

Degree: Bachelor of Science in Business Administration
Majors: Finance and Business Analytics
Minor: Computer Science
Cumulative GPA: 3.71/4.0

Relevant Coursework:
• Artificial Intelligence
• Robust Machine Learning
• Database Design & Implementation
• Statistical Inference
• Programming for Data Analytics (R Studio)
• Marketing Insights