Adam Rauckhorst is a Data Scientist with a background in business and operations. Prior to Metis, he oversaw operations for Greenback Expat Taxes, which earned a place on the Inc 5000 fastest growing companies during his tenure. He is passionate about using technology to achieve substantial business results. Adam earned a Bachelor's in Economics with high honors from the University of Texas at Austin.
Who Should I Follow? Recommending Twitch Streams Using Collaborative Filtering
In this project, Adam utilized multi-threading to gather data from the Twitch public API to create a collaborative filtering recommender system.
Allen Chen is an actuary and data scientist. His data science journey began with a degree in Statistics from UC Berkeley. He then enjoyed a career as an actuary working in the health insurance industry. Allen used his time at Metis to gain additional skills and experience in the full stack data science workflow. He is excited to continue on his data science journey, where there is always more learning ahead.
An inquisitive and explorative person by nature, Allen is interested in making solutions that improve people's lives. Experienced in biomedical research, social outreach, and startups/small businesses, he aims to make a lasting mark by helping others. A love for complex multidisciplinary problems and a wide ranging knowledge on various subjects, allows Allen to adapt easily to any situation.
Cross Media Recommendations
Big Data / Natural Language Processing / Recommender Systems
A recommender that can suggest books similar to a movie or tv show and vice versa, to expand the user's palette and prevent oversaturation of a specific medium
Air Force veteran with five years of data analysis and project management experience. Skilled in working with cross-functional teams, implementing Agile and Scrum methodologies, and leveraging data to make critical decisions in challenging environments. Recently transitioned out of the military and keen to leverage leadership/technical expertise.
Make Your Move: Image Detection with Neural Networks
Andrew Wu, is a recent graduate of UCSD with a degree in Chemical Engineering. His passions are coding, data science, data analytics. He has worked in the Biomedical field in the past as an analytical chemist and recently in a non-profit political organization as a data lead. In his free time, he likes to play music. Andrew played the violin for almost 10 years and also plays several other instruments as well. He is also very into fitness and works out regularly. Andrew is friendly and a people-oriented person. He has held leadership positions in student organizations and is good at bringing people together. Furthermore, he has a curious mind and considers himself a lifelong learner!
Wildventures: A Photo-Based Travel Recommender System
Classification / Neural Networks / Recommender Systems
Wildventures, a photo-based travel recommender system that makes travel destination recommendations based on the user's previous travel photos.
With a bachelors in Economics and History, Ben Shaiman is constantly driven by the desire to understand how the modern world works, but also just how it came into being. After graduating college Ben worked as an intern for the Freakonomics Radio podcast. With an academic background ranging from the history of antiquity to advanced statistics and economic theory, it was data where he found a calling. The unique challenge of communicating complexity to a lay audience, so critical in podcasting, drove Ben to pursue more technical experience.
Machine Learning Sommelier
Classification / Natural Language Processing / Natural Language Processing
Trained a neural network on over 250,000 wine reviews to classify wine varieties.
Binh is a data scientist transitioning from finance. He worked on the trading floor of D. E. Shaw & Co., a global investment and technology development firm, as a member of the portfolio financing team. His penchant for data analysis and visualizations led him to transition to data science. Outside of work, he is a community organizer in the queer and trans Asian Pacific Islander community. Binh graduated from Yale University with a B.A. in humanities.
A User-Empowered Anime Recommender System
Cloud Computing (AWS/Google Cloud) / Recommender Systems
Built an Anime recommender system based on how adventurous users want their recommendations to be.
Brian obtained his master's degree at the University of Illinois studying physical chemistry. After graduation, he spent the following two years working in process sciences at a pharmaceutical company where he supported the release of a new drug to market from the data end. He has worked on generating and analyzing data for nearly ten years and joined Metis looking to expand his tool kit for data-driven insights. Brian looks forward to taking the next step as a data scientist.
Tunable Movie Recommender
Natural Language Processing / Recommender Systems
Created a recommender system using three content filters and a collaborative filter. The recommender allows the user to tune recommendations based on the interest of the synopsis, theme, crew, and ratings.
Bryan graduated with an M.A. in Clinical Psychology from California State University, Northridge. Having strong interests in quantitative methods and consumer behavior, much of his research concerned predicting these behaviors and consumer-brand relationships. He pursued these interests professionally at a market research firm in Los Angeles. He worked with the Chief Innovation Officer, helping to develop insightful tools for use in company studies. In this role, he assisted with the design of a variety research and development studies, in addition to being the primary analyst. Growing more passionate about machine learning and programming, Bryan decided to attend Metis to strengthen these skills and pursue a career in data science.
Reddit on Relationships
Natural Language Processing / Neural Networks / Recommender Systems
Used a combination of Latent Dirichlet Allocation and BERT embeddings to create a content-based recommender system. Users can enter their relationship problem(s) into a Flask app and receive relationship advice posts from Reddit.
Recent graduate of Metis' Data Science program with advanced data analytics and machine learning skills. Prior to data science, 15+ years of experience in marketing, project management, research, and analysis spanning the non-profit, corporate, and government sectors.
Cianan recently graduated from UCLA with a BS in Biophysics. He has worked in several research labs ranging from Neurophysics to genetics to molecular biology. He also interned at a startup medical device company and helped with experiment design and overview. His passion for data science stems from these research experiences where his favorite part of the research was examining the collected data and drawing conclusions from the experiments. Cianan loves problem-solving and coming up with unique insights through thorough investigation. He works great in teams, loves learning, and is always looking to improve.
Wildfire Size Prediction
Anomaly Detection / Classification / Time Series/Forecasting
Cianan's project is about making a model that predicts how large a wildfire would be, if one were to start throughout California.
Cynthia holds a B.S. in Joint Economics and Math and a B.S. in Clinical Psychology with honors from University of California, San Diego. Her education in math and statistics and experience in psychological research in decision-making sparked her interest in data science. After she graduated from college, she earned a Master’s degree in Quantitative Methods in Social Sciences from Columbia University where she practiced data science in social science disciplines and harnessed programming skills in Python, R and SQL. Cynthia has a strong curiosity and is passionate about using data science to understand human behavior and offer solutions to the real-world challenges. With her knowledge in machine learning, statistical analysis and psychology, she’s excited to start her career as a data scientist.
Classical Music Generation with Deep Learning
Big Data / Cloud Computing (AWS/Google Cloud) / Neural Networks
This project uses deep learning to explore the intersection between arts and data science. It utilizes a model architecture with GRU and self-attention to generate classical piano music.
David is a Data Scientist based in New York City, working on several types of machine learning projects during his time at Metis. These projects cover linear regression, classification, natural language processing, and deep learning. David's goal as a data scientist is to ask the right questions and craft stories out of the raw data. His previous experiences include a variety of roles in the healthcare/technology industry at Komodo Health, Gerson Lehrman Group, and Columbia University Irving Medical Center. David graduated from Columbia University with a master's degree in bioethics, focusing on healthcare data and data privacy.
Interpreting Speech Emotion for the Hearing-Impaired
Some of the greatest challenges as a hearing impaired individual or in communicating with a hearing impaired individual arise when trying to start a conversation and when conveying emotion or intent. To address this challenge David built a convolutional neural network model to classify speech emotion and the gender of the speaker. He also created an app that processes input audio and notifies the user of the speaker's gender and the sentiment of the speech.
Edith is a creative thinker with a passion for learning new and diverse skills and information. A Bay Area native, she graduated from the University of New Mexico in Spring 2019 with a B.S. in Applied Mathematics. Edith has a wide variety of interests, ranging from astrophysics and chemistry (how the world works) to evolutionary anthropology and psychology (how people work). She is an adaptive problem solver with dogged persistence in the face of challenges, and thrives in the analytic, logic based environment of data science. In her free time, Edith enjoys creating beautiful things, such as embroidery, painting, weaving and crochet, and caring for her many houseplants.
Eliza graduated from Indiana University in 2017 with a B.A. in cognitive science. Since then she’s traveled extensively and worked in a variety of industries. Returning to her undergrad roots of science and math, Eliza began coding again and decided to pursue more technically focused roles. To close the gap, she applied to Metis and threw herself into four independently designed projects from Natural language processing of comedy podcasts to classifying chest X-rays using a neural network.
Emily graduated from the University of Illinois at Urbana-Champaign with a double major in Physics and Astronomy. She has experience in data processing and analysis while working with a wide range of astronomical data. After dipping into computer programming and machine learning concepts in undergrad, Emily developed an interest in leveraging data science for everyday technology and business applications. She is driven by her deep curiosity of the world and is quick to learn new things. Emily is optimistic about diving into the world of data and is most excited to apply her strong analytical skills to solve complex real-world problems.
Born and raised in the New Jersey/ New York City area, I have always been passionate about learning new things and solving big problems. I have a B.S. in Computer Science, a minor in Mathematics, and three years of experience teaching the two subjects at the high school level. Currently, I am excited about exploring the world of Data Science and progressing in my career by continuing to infuse my passions for Math and Computer Science. For fun, I love to write and record music, spend time playing through my board game collection, and find new ways to maintain my athleticism and good health.
Natural Language Processing / Neural Networks / Recommender Systems
This project offers guidance for people new to podcast listening. The Flask app stores user likes and uses a Word Embedding Neural Network to generate search results and recommendations from "This American Life."
A former History teacher, Esteban holds a B.A. in History from The City College of New York with several courses in economics and jazz studies. Prior to joining Metis he was a law student, but it was his fascination with business efficiency and problem solving that led him to pursue a more technically focused, full-time career in data science. Outside of data science Esteban enjoys playing tennis and has been a guitar tutor for 20 years.
Formerly an Agricultural Engineer, Faustina’s passion for data was sparked by her experience in soil analysis and statistics. Realizing that data-informed decision-making could, for example, increase crop yield in her native Argentina, she transitioned to full-time work in data science. Faustina works side-by-side with individuals and organizations to solve problems using Machine Learning solutions.
Controlled Chaos: A tool for Taming Digital Clutter
Clustering / Natural Language Processing / SQL
A web application running locally and automatically that organizes text files based on its content. Controlled Chaos maps and classifies files, builds file clusters, and indexes them for easy application-based access.
Frederick Lam has a B.A. from Roosevelt University, majoring in Actuarial Science and a minor in Finance. During his studies, he found interest in the data manipulation and analysis side of Actuarial Science and it inspired him to pursue more data-focused roles after graduation, where he found out about the Data Science field. After searching for a year, he was pointed to the Metis Data Science bootcamp by a fellow Data Scientist. As a student at Metis, he took the opportunity to combine his passions, such as anime and video games, with complex machine learning algorithms and modeling. As a result, he created an anime recommender, based on the collaborative filtering technique, as his final project at Metis. His motivation for this recommender was not only to present the skills he'd learned at Metis but also to showcase how data science can be applicable to a general level of interaction and understanding.
Satisfying Your Anime Needs
This project is a take on Collaborative Filtering Recommender systems and how it works on a simple user-based level. This project also focuses on taking data from a reputable source, MyAnimeList, and creating an interactive recommender app using it.
Gio DeLisa has a BA in Economics with a minor in web development from New York University. Prior to Metis he helped to build and grow a small New York based apparel brand with a focus on local manufacturing. Always interested in learning, Gio is looking forward to being part of the data science community and its continual developments.
Forecasting for Fantasy Birding
Big Data / Neural Networks / Time Series/Forecasting
Developed convolutional sequence-to-sequence neural network to forecast expected amout of recorded bird ovservations in over 4,000 east coast locations. Used dask to process 100M+ bird obsrevations. Used the H3 library for geospacial indexing.
Recently graduated with a PhD in Pure Mathematics from the University of Illinois at Chicago. While in graduate school, spent time working as a Teaching Assistant and a Research Assistant. Pursuing a career in Data Science, as it would provide the perfect opportunity to apply the research/problem solving skills cultivated in graduate school, to relevant problems in the real world.
Detecting Ransomware Payments
Big Data / Classification / Cloud Computing (AWS/Google Cloud)
Created a model to detect Ransomware payments using publicly available data on bitcoin transactions. Used KMeans clustering to identify two distinct waves of Ransomware and trained models to detect each wave.
Harry John Shephard is a theatre nerd and board game geek who has worked in the nonprofit sector for more than a decade and has seen many opportunities for institutions to use data science but haven't incorporated it into their services and organizational processes. Transitioning into data, he is eager to work in an business with a strong data culture where he can learn from experienced and skilled leaders who are enthusiastic about using data in creative and surprising ways. He is particularly interested in NLP focused projects. If you know of a nonprofit, community organization, or small independent business that could use some pro-bono data work, please don't hesitate to pass on his information to them.
The Language of Leftovers: Exploring Communication In Eating Disorder in Online Communities
Classification / Cloud Computing (AWS/Google Cloud) / Natural Language Processing
Explored trends and looked for connections in the words used between different Reddit sub-forums within communities directly or indirectly related to eating disorders.
Isaac graduated from UT Austin with a BBA in Management Information Systems. Prior to joining Metis, Isaac was a data analyst at United Airlines, where he oversaw Voices, and internal app for flight attendants to report operational issues. Isaac built dashboards in Palantir for various departments to identify and track key trending issues. Utilizing many different data science concepts, he performed logistic regression to predict flight attendant churn and natural language processing to perform sentiment analysis on newly offered United in-flight products. Isaac also has Tableau experience, building an end to end fully interactive, functional dashboard for operational teams to track productivity. Now, Isaac is looking to build off of his strong analytical background and pursue a career in data science.
Generating Lo-Fi Music with Neural Networks
Big Data / Cloud Computing (AWS/Google Cloud) / Neural Networks
Utilizing deep learning, applied knowledge of neural networks, specifically LSTM and self-attention to generate Lo-Fi music (chill, distorted short jazz loops with a beat).
Before joining Metis, Jacky was a Life Science Product Specialist at Quartzy. He assisted customers with technical questions about products, curated a database of scientifically equivalent product recommendations, and maintained the integrity and clarity of data for a catalog of over 10 million life science products. He wants to combine his interests in operations, logistics, and technical support with his skills in data analytics, project design, and programming in his pursuit of a full-time career in data science. Jacky holds a Bachelor of Science in Molecular Biology from UC San Diego.
Exploring U.S. National Parks
Neural Networks / Recommender Systems / Time Series/Forecasting
Jacky used a SARIMA time series model to predict monthly visitor counts for national parks. Additionally, he built a content based national park recommender using a topic activity cosine similarity matrix.
Jimmy graduated from Clark Atlanta University obtaining a B.A. in Finance & Computer Science. The find a way or make one approach, contributes to his passion for providing voice to data through storytelling that cultivates action. Prior to being a Metis alumnus, his career spans across multiple industries such as Banking, Investment Management, and Healthcare. These acquired skill sets positions him to close gaps amongst technical & non-technical stakeholders. Overall, the enthusiasm about embedding his entrepreneurial spirit to create innovative solutions as a data scientist is invaluable.
WHAT ARE THOSE!?
Classification / Neural Networks / Recommender Systems
Developed a sneaker recommendation system. Pictures were leveraged to train a convolutional neural network (VGG16) which identified & classified sneakers by brand.
Recent Data Scientist bootcamp graduate with design background in architecture; fascinated by the possibilities of machine learning and big data. Prior to the Data Science Bootcamp, John was the Technical Director and a Project Manager at a creative/design oriented architecture + real-estate development practice.
Architectural Form Generation using LiDAR Surveys of Architectural Interiors
Jon is a long-time Statistician and in recent years has transitioned into Data Science. He has an MS in OR & Statistics and has worked in both academia and business settings. The main tools he has used throughout his career are SAS, Simca and R for analysis and visualization. His main focus was design of experiments, predictive models, statistical process control and multivariate analysis. Most recently Jon was consulting through is own business, but faced some challenges in maintaining his client base when the coronavirus pandemic hit. Jon wanted to attend Metis to expand his analytical toolkit and felt like doing so during the pandemic was good timing. He feels his skills using Python and SQL, as well as AWS have improved greatly. He now has a more robust skillset when it comes to machine learning and natural language processing and is ready to use these skills to contribute to an organization as a data scientist.
Predicting Loan Profit Margin
Regression / SQL / Tableau/Dashboards
A finance company is concerned that their profit margins are declining. They want to create a prediction of the profit margin target using various loan and customer features.
With over a decade of experience in the management consulting industry, Jonathan primarily focused on Business Intelligence and Visualization and is looking to leverage those skills with full data science tool kit. Jonathan is an avid traveler of the world, a car enthusiast, a husband and a father of 3.
Former graduate of UC Davis with a degree in Economics and Specialty in data analytics & economic analysis. After graduating, Joseph worked 2 years for YouGov as a research analyst, before ultimately wanting to take his skillset further to become a data scientist and enrolling in Metis.
COVID-19's Impact on the 2020 Election
Classification / Regression / Time Series/Forecasting
Looking into COVID-19, polling, and past voting datasets to make findings about how COVID-19 has affected the 2020 political landscape.
Jung-a Kim is a data scientist with background in statistics. Previously, she worked as a contractor in Intuit where she demonstrated benchmarking a pre-published deep kernel learning model and she worked as a web-programmer in a database company. Her problem-solving skills have been endorsed by colleagues and she has a strong leadership in projects. She earned a M.S. in Statistics from San Jose State University and two years of coursework in Computer Science at De Anza College.
Review-Based Search Engine
Natural Language Processing / Neural Networks / Recommender Systems
A review-based search engine outputs optimal products whose reviews match key qualities from queries. It computes 'Positive Similarity score' which is a combination of positive sentiment score, similarity matrix, and weights on the keywords.
Justin is a Bay Area native with a B.S in Accounting from Santa Clara University. His previous role was in Non-Profit Finance at the Kenneth Rainin Foundation. During that time, he was tasked to create various dashboards that were used for investment decisions. Most importantly, he scrutinized and ensured the accuracy of any report or dashboard so that the investment team can be 100% confident in their decision making. He's a highly focused individual, making sure there's no stones left unturned on whatever task he seeks out to finish. On his free time he enjoys golfing!
Classification / Neural Networks
Using neural networks for handgun detection within a video.
Kaitlin is a researcher with 5+ years experience in the space of genomics, DNA sequencing, and big data analysis. Throughout her experience, she's developed a passion for parsing through large datasets and adding interpretability to a seemingly boundless space.
More recently, Kaitlin has pursued a career pivot into the space of data science, prompted by an interest to combine deep learning and genomics data. Her strongest interest is in the space of natural language processing, specifically in the spaces of conversational AI and the applications of NLP on non-traditional vocabularies.
Kelsey graduated from the University of California, Davis in 2017. While pursuing his degree in linguistics, he developed an initial interest in data science through coursework in computational linguistics and NLP. After gaining experience in political research, he spent the past two years teaching high school speech and debate in the San Francisco Bay Area. But, despite a passion for helping his students find success, a growing desire to further his journey into the world of data and A.I. drove him to enroll at Metis. By combining a strong foundation in machine learning with his background in education and qualitative research, he seeks to bring data-driven perspectives to big-picture strategic thinking.
AniMaker: AI-Generated Story Concepts from Anime Plot Synopsis
Natural Language Processing / Neural Networks / Regression
Generation and curation of unique story concepts through training GPT-2 and regression models on Anime series plot synopsis and community ratings.
Krystyna Metcalf holds a Bachelor of Science in Finance and International Management from Boston University. Prior to joining Metis, she was a Senior Equity Research Associate for nearly seven years covering the Household & Personal Care industry. Here she developed a passion for analyzing big data to identify patterns and predict trends. Outside of work, Krystyna is a passionate equestrian who has successfully competed on the international level.
Winston Churchill: A Study in Oratory Excellency
Cloud Computing (AWS/Google Cloud) / Natural Language Processing / Tableau/Dashboards
Through the use of Natural Language Processing, Krystyna analyzed Winston Churchill's speeches to discover major themes and how they changed over the years. She then compared the topics to current events at the time.
Lauren was drawn to data during her time as a massage therapist and sole proprietor of her business. Realizing the value in working with the health care data created at her work she began studying data analysis in R. Learning fundamentals of statistics became an interest of hers at the San Francisco City College. Python is now the language she is most fluent in in her work as a data scientist. In her free time she enjoys pattern making and bird watching. She appreciates the chances to observe unseen aspects of daily life during her time with data.
Bird Counts of Costa Rican
Big Data / SQL / Time Series/Forecasting
Bird Counts of Costa Rica are forecasted using Vector Auto Regression algorithm. These predictions are made from citizen science project eBird data. Google Cloud Platform's Spark Data Proc Cluster facilitates cloud computing needs.
After graduating from Colorado State University in 2018 with degrees in Microbiology and Spanish, Lindsay worked as a Microbiologist and shift lead, training new employees and assuring quality and timely results to customers. Considering a career in Epidemiology, she became fascinated by the integral part data played in highlighting how disease outbreaks could be mitigated and/or prevented. This passion, along with a lifelong admiration for statistics and data visualization, led Lindsay to embark upon her journey to data science. For her final project, she used NLP and feature engineering techniques to create an interactive recommendation system for hotels in Puerto Vallarta, Mexico.
Hotel Recommender for Puerto Vallarta, Mexico
Natural Language Processing / Recommender Systems / Tableau/Dashboards
Combined NLP and feature engineering techniques to create a search engine that applies transfer learning from word embeddings to recommend the most (cosine) similar hotel(s) for a user based on his/her text input and selection of features.
Lisa earned a B.S. in Mathematics and Psychology from St. Lawrence University and a M.A.T. in Secondary Mathematics Education from National Louis University. A first generation college graduate, Lisa joined Teach For America in Chicago following graduation, and was an award winning high school and middle school math teacher on the south and west sides of Chicago. She spent the last 4 years as an administrator in a middle school in Chicago, using data to drive transformational change in student outcomes. Lisa is passionate about using data to tell stories, drive results, and create lasting change.
Forecasting Divvy Bike Share Demand During Covid-19
SQL / Tableau/Dashboards / Time Series/Forecasting
Forecasted Divvy bike share demand for the remainder of 2020 using data from the City of Chicago from January 2017-August 2020, along with information from Chicago’s phased Covid-19 shutdown and reopening. Developed a Tableau Story to visualize Divvy bike share demand as a function of Chicago’s reopening.
Louis has never had a problem asking questions- his curiosity seemingly knows no bounds- and data science allows him to put that curiosity into practice. From coming up with questions to finding insights in the results, there is no shortage of interesting discoveries to be made along the way. In a field that is continuously expanding, Louis finds himself well-prepared and excited to get started.
Visualizing Policing in Chicago
Classification / Tableau/Dashboards
Can we predict if a reported crime in Chicago will result in an arrest based on surrounding area of the crime?
Graduated from Colgate University with a degree in mathematical economics in 2018. Then spent a year as an AmeriCorps volunteer, tutoring middle school students in math at a charter school in Newark, NJ.
Ski Resort Recommender
Recommender Systems / Tableau/Dashboards
Created a web application to recommend ski resorts to new or experienced skiers.
Originally from Spain, Manuel has lived in New York for the last 6 years working in a neuroscience research laboratory at SUNY Downstate studying perception. His background pivots between Optometry and Physics. Before moving to the US he ran his own business, but always had a special interest in science, which is what brought him to the US. Now, looking for a career change, data science is the perfect match, a balance between continuous learning, business perspective, and applying his experience to new challenges.
Big Data / Cloud Computing (AWS/Google Cloud) / SQL
Inspired by the business traveler, used Neo4j to gain a deeper understanding of the US airports network structure applying different centrality, clustering and path finding algorithms.
Maryam has a strong background in mathematics and statistics and holds an M.S. in Applied Mathematics (Actuarial Science) from the University of Illinois at Urbana Champaign. Following grad school, she joined PricewaterhouseCoopers (PwC) in the insurance and M&A industry. She worked as a manager of Actuarial Science, supervised her teams to perform actuarial valuations, and consulted clients on pension plan designs and risk mitigation, often tackling multiple projects and deadlines at the same time. She enjoyed the heavily analytical and data-oriented nature of her work and was often using SAS, Alteryx, and Tableau. Following that passion for data analytics and machine learning, she decided to pursue a career in data science.
Generated audio captions for images using Neural Networks to help people with visual impairments. Analyzed images using CNN and text data using LSTM, and then converted captions to audio using WaveNet.
Max Currier is a data scientist, analyst and experienced manager of data-driven teams. Previously, he served as a Senior Project Manager at Gartner and as Analytics Manager at China Beige Book International where he managed large teams of analysts and data collection specialists through numerous large-scale, fast-paced projects that were central to their firms’ core business. Max’s natural affinity for problem solving has lead him from a BA in Chinese language to working in the data and analytics space, and now ultimately to pursuing a career in data science.
SoftCopy: Image-based Book Recommendation
Natural Language Processing / Neural Networks
SoftCopy is an app that generates book suggestions based off images of a user’s book collection by performing optical character recognition and employing a collaborative filtering recommender system.
Michael graduated from the University of Maryland, College Park with a degree in Economics. After graduating, he worked for a public health organization for five years where his most recent role was a production coordinator. In his role, he got exposed to learning and using data visualization with Tableau. By learning about data visualization, it peaked his interest in data science where he wanted to learn more about analyzing data and machine learning. For his final project, he used deep learning to see if he could detect anger in text.
Angry Text Detector
Classification / Natural Language Processing / Neural Networks
Michael set out to analyze the sentiment of text. He wanted to predict whether a text was angry or not. He did topic modeling, created a recurrent neural network model, and created a Streamlit app.
Nick is motivated, competent, and hard working and holds a bachelors in Anthropology from UCSB. Nick believes that his background as a technical writer, where he consulted with engineers and a high school tutor, where he was recognized for reaching difficult students provides him with a rich experience where he can contribute meaningfully to a team.
Image Creation with Generative Adversarial Networks
3 web apps for manipulating images in different ways - style transfer, face swap, and latent representation blending.
Before enrolling in Metis, Paul worked in the commercial real estate industry for 15 years performing transaction and asset management advisory services for institutional investors. As he started dealing with larger sets of data on a daily basis, he felt that he needed more than VBA in MS Excel to get the most out of his analyses. Given his education in Electrical Engineering, moving into the data science field seemed to be next sensible step.
Nudge Your Run Pace with Music
SQL / Tableau/Dashboards / Time Series/Forecasting
Coded a model that predicts a runner's pace during various race segments and adjusts the runner's music to either dial up or pull back their effort.
Preston attended Washington University in St. Louis where he received his bachelors of marketing and psychology in 2018. After graduating, Preston went on to work as a Business Analyst and managing accounts at Walmart.com. He enjoyed working with technology and the analytical aspects to his work and after two years in his role he decided he wanted to challenge himself further and expand his analytical toolkit. He decided to attend Metis to gain stronger analysis and decision making skills. As a Metis graduate Preson is actively looking for new opportunities and is especially drawn to Business Intelligence in the finance or civic engagement space.
Creating a Forecasting Engine for Chicago's 311 Services
Clustering / Recommender Systems / Time Series/Forecasting
Analyzed Chicago 311 data and found that service times are too long and vary too much across neighborhoods. To address the issue Preston built a forecasting engine to allow the city to plan resources accordingly and reduce service time.
Ramon served in US Army for 10 years as infantry squad leader and special operations engineer. After his discharge he received Bachelor's degree in Robotics/Mechatronic Engineering and worked as Datacenter Technician, Manufacturing Engineer and a technical cofounder for Aviation eLearning App. With a strong engineering background, Ramon is leveraging his previous experience with data science.
Automatic Mask and Temp Station
Anomaly Detection / Classification
Automatic contact free station that detects a person's temperature and if they are wearing a mask.
Raymond graduated with a BA in Computer Science from UC Berkeley. Prior to Metis, he was a quantitative researcher at Old Mission Capital where he led the research efforts on their options market making desk and managed risk across two low-latency strategies. At Berkeley, he has done research under several professors in computational game theory and latency arbitrage. He co-founded Fluint while in school, an angel-backed startup providing an online marketplace for peer-to-peer foreign currency exchange. Raymond enjoys taking on challenges in the areas of quantitative finance, behavioral sciences, computer vision, and game theory. Outside of work hours, he’s an avid poker player and travel photographer. Ask him for international travel and food recommendations or tips on collecting credit card bonuses.
AttackGAN: Adversarial Attacks Using GANs
Classification / Neural Networks
Researched new methods to generate adversarial examples using generative adversarial nets to fool deep image classifiers.
Rudy holds a BS in Finance and has three years of experience in financial risk and regulation. His curiosity for combing through Big Data and delivering innovative solutions and strategies combined with his analytical and quantitative skill sets is the perfect match for a career in data science. With data processing and visualization tools , Rudy looks to help businesses drive strong results-oriented solutions.
Ryan Richbourg earned a BBA from Baylor University with dual majors in Entrepreneurship and Supply Chain Management. Prior to Metis, he worked for three years in sales and marketing roles across several industries including research technology, outdoor recreation, and e-commerce. He transitioned to data science because he loves continuously learning, solving unique challenges, and using programming to extract insights. Ryan's mission as a data scientist is to help organizations translate their business problems into technical solutions using data. His final project at Metis uses natural language processing techniques to analyze topics and create a post recommender for Seth Godin's blog on marketing and growth.
Text Analysis on Seth Godin's Marketing Blog
Clustering / Natural Language Processing / Recommender Systems
Scraped 7500 posts from Seth Godin's marketing blog and clustered them into 7 main topics using NLP topic modeling. Additionally created a content-similarity based recommender system inside an interactive web app to suggest further reading.
Sam began his career as a mechanical/controls/robotics engineer in the aerospace manufacturing industry. He encountered data analysis and statistics in the context of measurement, inspection, alignment, compensation, and qualification of parts and machinery and developed an interest in building further machine learning and programming skills at Metis to take part in the ongoing new industrial revolution. His areas of interest include manufacturing, industry, mechanical design, robotics, signal processing, sustainability, and environmental policy.
Sasha Prokhorova holds a B.A. in Linguistics from Smolensk State University and a B.S. in Electrical Engineering from San Francisco State University. Her love for programming, problem-solving and data visualization led her to pursue a full-time career in Data Science. For her final project at Metis she analyzed data obtained through The New York Times Archive API to derive insights about gender representation in the media throughout the past 70 years. She hosts a podcast on YouTube called SashaTalksTech, where she interviews a variety of data science professionals from all over the world, sharing their knowledge, expertise and success.
Women Through the Lens of The New York Times
Natural Language Processing / Tableau/Dashboards
Data journalism research on gender equality and representation in The New York Times. Data collection through the Archive API for 1950 - 2020, topic modeling using SpaCy, gensim and Latent Dirichlet Allocation technique. Frequent term visualization based on the keywords and the timeframe. Sentiment analysis using NLTK Vader.
Sean graduated from University at Buffalo with a Bachelor's in English Literature. His data science skill set was developed at Metis where he defined, executed and presented five full scale projects based on regression, classification, and unsupervised learning techniques. Prior to Metis, Sean worked in the hospitality industry and wrote creatively. He looks forward to leveraging his passion for storytelling in a data driven role.
Tremendous Patterns in Speech: An analysis of the Trump Administration
Natural Language Processing / Tableau/Dashboards / Time Series/Forecasting
An analysis of the President and his press secretaries using transcripts from the White House Press Briefing and Press Releases.
Sophia holds a Masters degree in Accounting Science and Finance from the University of Illinois at Urbana-Champaign. Prior to joining Metis, she spent years at Ernst & Young as a Consultant and a Project Manager before transitioning to Merrill Lynch where she worked closely with the executives to deploy market wide strategic proposals. After spending nearly a decade in the financial services industry, Sophia wanted a new challenge where she could utilize her ability to capture the big picture concepts and her analytical nature. A career in data science was the perfect progression. Through Metis, she learned new tools and gained hands on experience solving real world problems.
Supervised Image Classification: Let's Get Toned!
Big Data / Classification / Neural Networks
Created a streamlit app designed to detect a person’s skin undertone using a supervised image classifier built with convolutional neural network (CNN) and transfer learning.
I am an engineer by background and I have spent the past ten year in data analysis related roles in the Oil and Gas industry. I am passionate about new technology and data, so I decided to expand my toolkit and transition into Data Science full time. I am looking forward to applying my new skills and continuing to grow as a data professional.
Field Hockey Player Tracker
Big Data / Classification / Neural Networks
Using computer vision techniques to track field hockey players from video footage, determine their team from the color of their jersey and projecting locations onto a 2-D map of the field
After graduating from Wheaton College with a Physics major and Mathematics minor, Stephen has spent the past seven years overseas working as a professional educator of both Science and English, accumulating both a Teaching Certification and CELTA in the process. Since encountering computer programming in early 2020, he has become enthralled with the field and the abstraction and critical thinking demanded of it. For his final project, Stephen manipulates an autoencoder neural network in order to generate new and unique works of art. In his personal life, Stephen is an avid traveller and adventure seeker as well as a loving husband and father.
Machine Art: Generating images with a Convolutional Neural Net
Clustering / Neural Networks / Recommender Systems
Convolutional Neural Nets are frequently used to process and classify images. This project takes that function one step forward by using a neural net to create new and unique artistic images.
Stephen is a graduate of the University of California - Berkeley, where he obtained a B.S. in Mechanical & Nuclear Engineering. After graduating in 2014, he spent a couple of years at two Bay Area startups in the green-tech space as a Software Engineer. Most recently he worked as an engineer at a utility-scale solar panel manufacturer, researching and developing algorithms to predict the performance of solar power plants, analyzing power plant performance via high resolution time series data, and developing software in Python to enhance the company's analytics tools. His fascination with machine learning and artificial intelligence grew over the past few years until he decided to enroll at Metis. Stephen is now based in Denver, Colorado. While he is particularly interested in finding ways to apply machine learning to renewable energy and music technology, he is open to a wide range of opportunities.
Object Detection for Autonomous Snow Grooming Applications
Big Data / Classification / Neural Networks
A proof of concept for autonomous snow grooming vehicles at ski resorts. Trained a Faster R-CNN object detection model in PyTorch to detect/classify obstacles in dash cam footage from vehicles moving through ski resort.
TJ (he/him) has developed an interest in how we look for patterns and construct meaning, based on his studies of Linguistics and Theatre at Rice University. He has worked in the theatre industry for the past seven years in New York; mainly backstage and on the design teams for immersive experiences like Third Rail Projects’ Then She Fell and Behind the City, as well as Broadway and off-Broadway shows like Moulin Rouge!, The Band’s Visit, The Lucky Ones, and KPOP. TJ sees data science as a broadening of the scope of his analytic endeavors beyond the worlds of linguistics and theatre. Well-designed visualizations and engaging topics are important to his projects. He looks forward to learning more and continuing his data science journey.
An Ode to Data Science: Generating Poetry with Neural Networks
Trained Markov chains via markovify, an LSTM model, and GPT-2 (124M) on three poetry datasets to study the effects of text size and homogeneity on text generation and the strengths and weaknesses of each model.
I'm from Chicago, where I'm also currently based. Before coming to Metis, I was a research student, most recently using neuroimaging to study the link between temporal lobe epilepsy and mood disorders. I graduated from Notre Dame in 2015, where I studied biology, and I earned my master's from Rush University in 2018. While neuroscience remains a topic of great interest to me, the data science skills I've developed at Metis have led me to expand my horizons to new fields. Apart from data science, I enjoy board games, live music, reading, cooking, and trivia. I love animals, especially dogs and birds.
Bird Image Classification Using CNNs and Transfer Learning
Classification / Neural Networks
Using convolutional neural networks and transfer learning, created a program to identify the species of a bird from a photo and then sort that photo based on the species.
Daniel graduated from Juniata College with a BS in Biology and finished a Post-Baccalaureate in cell biology at the University of Pennsylvania. His drive for data science stems from performing data analysis while working as a research associate at Unity Biotechnology. As a researcher, Daniel was tasked to handle data derived from experiments, draw conclusions and present them to the team. Daniel enjoys solving problems, troubleshooting issues, and coming up with solutions in a timely manner. He thrives in team settings, and his skill in effectively communicating with others is what drives his ability to solve a variety of problems. For his final project, Daniel built a neural network capable of identifying the breed of dog from images of dogs.
Dog Breed Recognition System
Using neural networks to build a model that is capable of identifying the dog breed from an image of a dog.
Daria holds a degree in Economics from Siberian Federal University. She previously worked as a Credit Analyst at Sberbank, Russia’s largest bank. During her time there, she worked on approving ~$100M of corporate loans. Additionally, Daria worked as a data and analytics intern for rmdy.com, a venture backed ecommerce startup, which ignited her passion about data and its ability to help drive key business decisions. While at Metis she has used this motivation to create projects with specific business insights – including helping an emerging bank make better loan decisions and improving an online clothing brand make actionable insights form their reviews. In her free time, she loves hiking (and plans to go to every National Park), running, and travelling.
Optimizing Online Advertisements with Convolutional Neural Networks
Classification / Natural Language Processing / SQL
Helping companies improve ad efficiency by matching their advertisements to website content using a convolutional neural network model to classify images and choose the closest to reflect what is on each webpage.
Dmitri graduated with a M.S. in Physics from Moscow University and MBA in Finance from St. John’s University. Passionate about applying scientific methods and creating methodology to systematically improve decision making, Dmitri spent last 10 years in commercial real estate focusing on market analysis, consulting, investment analysis, and cash flow projections. Strong believer in technology, Dmitri utilized his background in solving business problems or accomplishing desired business outcomes. Through that interest in data science insights, Dmitri decided to pivot his career and joined Metis.
Elena holds three Master degrees in International Relations, Economics, and most recently completed an Applied Analytics masters program from Columbia University. Before making a full pivot towards data science, Elena worked at Microsoft for 10 years and was tasked with developing cloud business and transforming partner ecosystem across 33 countries in Central and Eastern Europe. In addition, she worked with key enterprise clients and leading sales teams. Elena is involved in educational programs at Columbia University and serves as an associate and an instructor. Apart from data science, she is interested in studying human psychology, languages, books, and visual arts. Elena Dubova resides in New York.
Blogger Boost: Know Yourself and Your Community
Classification / Clustering / Natural Language Processing
A web application for bloggers with three functional blocks: topic-based community visualizer, smart dictionary based on word embeddings specific to the community, and a set of dashboards that analyze emotional profile of one's writing.
A data scientist and Stanford graduate focused on leveraging data to help guide business decisions. A quick learner who utilizes a balance of analytical interpersonal and skills to problem solve and consistently deliver on goals.
Predicting the Winner of Counter Strike Rounds
Time Series/Forecasting / Natural Language Processing / SQL
Used Random Forest and extensive feature engineering to correctly classify the winner of CS:GO rounds.
Fong Wa brings to data science an enterprising sense for how to use data to navigate the business landscape, within companies both large and small. She is respected for her ability to capture the big picture alongside the details and mobilize stakeholders to engage with the mission at hand. Her career prior to Metis includes meaningful roles in corporate and entrepreneurial organizations including financial services and conversational AI.
In her professional and personal life, Fong Wa is a charismatic go-to person for her colleagues, friends and family when seeking practical answers and tackling challenging problems. She holds an MBA from the University of Chicago and a BA with honors from the University of Pennsylvania.
Fong Wa C.
Clairvoyant Clustering of Consumer Complaints
Classification / Clustering / Natural Language Processing
A multi-faceted approach to analyzing text and clustering companies identified in consumer complaints filed with the Consumer Financial Protection Bureau. Tools and techniques include NLP (SpaCy and CorEx) and unsupervised machine learning (K-Means and PCA).
Jeff is a military combat veteran with a background in financial services. He holds an MBA from Baruch College and a B.A in Psychology from Binghamton University. His love for data science stems from his time deployed overseas where he used data analysis and problem-solving to recognize enemy attack patterns and predict their movement. In his most recent role, Jeff led business advisory projects at EY, using statistics to help multinational banks optimize their research and development credit. Along with his passion for data analysis, Jeff voluntarily prepares tax returns for low income families. He also enjoys the outdoors and has summited the tallest mountains in the US. For his final project, Jeff is building a neural network to identify automobiles to facilitate a safer autonomous driven world.
Image Classification with Convolutional Neural Networks
Big Data / Classification / Natural Language Processing
Created convolutional neural network model to classify images with high accuracy and facilitate autonomous vehicle operations.
Before studying data science, James studied physics at Harvey Mudd College, worked in biomedical research, tutored math and science, and worked as a software engineer designing big data tools, for applications such as product recommendation for retail, and entity recognition for the financial sector. He's excited to leverage his math and science background to help businesses make smart decisions as a data scientist.
Recognizing Game States in Super Smash Bros. Melee
Recommender Systems / Cloud Computing (AWS/Google Cloud) / Big Data
Trained a neural net to recognize the state of a character in a fighting game from still frames of videos of matches.
John is passionate about using data to help inform better decision making, which was the primary driver for his decision to join Metis. John's work experience includes cleaning, analyzing and generating insights from a variety of datasets for several different stakeholder groups within his previous investment bank firm. Through Metis, John dedicated his time to working on several machine learning projects, leveraging new tools along the way. Prior to his career, John studied at Stony Brook University in NY where he obtained a combined Bachelors/Masters Degree in Applied Math and Statistics. Outside of work, John enjoys fitness, travel and camping. He's travelled to 20 countries within a span of four years!
Forecasting Citi Bike Demand
Classification / Natural Language Processing / Regression
Forecasted the daily demand for Citi Bike usage with a 365-day horizon using Facebook Prophet and WaveNet models. Developed a web app with Dash to visualize forecasted results for all stations with map navigation.
Joshua is a life long lover of learning, and he greatly enjoys the deeper understanding of complex systems. Joshua is interested in the practical application of data science to business decision-making. He practiced law happily in Florida for some years but decided he wanted to begin pursuing data analytics by pursuing an MBA with a Business Analytics Concentration. Subsequently, Joshua continued to teach himself independently online, and now he has completed the Data Science Boot Camp at Metis. Joshua would like to be a balanced contributor in a collaborative team environment, and he hopes to offer a bit of everything, i.e. data gathering and exploration, analysis and interpretation, and communication and insight. Hopefully, his non-traditional background in getting here will complement your team and provide a helpful perspective.
Mining the Cart: Customer Segmentation Analysis
Time Series/Forecasting / Classification / SQL
The customer segmentation of retail consumers via the K-means algorithm using a large transactional retail dataset, and the interpretation of these clusters characteristics and implications for the business.
Kayla is a Brooklyn native and has an innate curiosity and passion for getting to the crux of how things work. That passion is what attracted Kayla to neuroscience and ultimately obtaining her masters in clinical neuroscience in London. After graduation, she left the world of academia and landed her first job in sales strategy for a life sciences consultancy. In that role, Kayla discovered how much she enjoyed using her creativity to solve business problems analytically. Early on in her career, Kayla's strong communication skills were valuable when it came to bridging the gap between technical and non-technical stakeholders, which she continues to hone.
MedRec: A Drug Recommender System
Natural Language Processing / Recommender Systems / Tableau/Dashboards
Drug recommender system that uses NLP techniques and NMF. User enters symptoms in conversational English as input. Recommendation system gives symptom-relevant, highly rated drug recommendations along with relevant positive reviews as output.
Luke is a native Californian and has lived most of his life in Silicon Valley. He served a few years in the US Army and enjoyed jumping out of perfectly good airplanes. Luke studied math as an undergrad and bioinformatics as a graduate student. He's always been interested in math, science, and technology, so he looks forward to working as a data scientist as it combines all those disciplines.
Human Protein Image Classification
Regression / Classification / Natural Language Processing
The goal of this project was to develop a binary classification model to distinguish between proteins in microscope images.
Max attended the University of California Santa Barbara where he obtained a Bachelor's degree in Physics. After college he worked at a startup, Wyatt Aerosol Systems, developing a system for aerosol particle characterization. Max moved to Chicago with his significant other where he has been continuing to expand his knowledge in data science and analytics. In addition to competing and finding joy in Kaggle competitions, Max decided to enroll at Metis to advance his knowledge in data science. He looks forward to pursuing a career in the analytical space.
Custom PC Build Recommender
Clustering / Recommender Systems
Developed a recommender system for custom PC-builds using data scraped from PcPartPicker.com.
Raised in Minnesota, educated in Chicago, Mitch now calls the City of Big Shoulders his home. It's been a wonderful path to the present, as his previous experience has developed a myriad of skills to bring him to data science. From account management, medical education, running a marketing department and video marketing production has left him with a humming curiosity that previously had nowhere to go. Data science allows Mitch to understand problems from the ground up and lets him communicate the story with gusto. Mitch relishes working within teams and can't wait to unveil more data stories.
The "Great Deliberators": Senate Speeches on Education, Banking & Healthcare
Natural Language Processing / Tableau/Dashboards / SQL
Exploring how the Senate has changed over the last four decades through NLP analysis of the topics covered during Senate floor speeches on education, healthcare and banking from 1980-2016.
Experienced data scientist with a background in Finance. Interested in all things data, from building an end to end data pipeline to ETL, data exploration and analysis. Particularly interested in gaining insight from data by building and deploying machine learning models and strategizing based on those insights.
Paul graduated from NYU with a degree in Economic Policy. He was a Data Analyst at Digitals where he worked with Comcast Residential and Comcast Business. Prior to Digitas, he served as a Business/Analytics Analyst for Mansueto Ventures. Paul is excited about leveraging his past experience towards a full-time career in data science.
Medicine Review Classification
Classification / Natural Language Processing / Tableau/Dashboards
Built a text (sentiment) classification model and applied NLP topic modeling techniques using medicine reviews.
A data science his entire career, Paul is a geologist by training with a thorough background in physics, chemistry, and mathematics. He graduated summa cum laude from Washington University in St. Louis and obtained a PhD from the University of Notre Dame with a dissertation focused on analyzing crystal structures from a large database. Paul worked for the State of Indiana, obtained his professional license, then returned to research and completed two postdocs. For the second postdoc Paul created a complex regression model to understand meteorite mineral chemistry. After teaching for several years, Paul decided to change course and became an independent consultant while working to transition to data science full-time.
Trouble in Paradise: Hawai'i in true and false colors
Big Data / Clustering / Regression
Principal component analysis, false color imaging, unsupervised spectral classification, and landform cluster analysis of multispectral image data from Hawai'i Island.
Terry is an experienced professional in the areas of finance and education. He has a variety of interests, and is especially drawn to the tech space. Terry is a hands-on person by nature; appreciates knowing how things work from the ground up. One of his strongest skills is the value he places on collaboration. Terry is not afraid to go it alone, but most outstanding achievements are the result of many contributors. Terry makes a conscious effort to include everyone in the pursuit of excellence.
Predicting Pediatric Pneumonia using Machine Learning
Classification / Cloud Computing (AWS/Google Cloud) / Big Data
For this project, Terry built three models (RandomForest, XGBoost, and a 3-layer CNN) which predict if a pediatric chest x-ray is normal or indicates a pneumonia infection.
Xin Cheng holds a MSc in Biological engineering from the University of Georgia. Prior to joining Metis he was a scientist at Synthego doing gene editing in production group. His love of data mining, problem-solving, and programming led him to pursue a full-time career in data science. For his final project he is analyzing depressive posts on social media and made a joke generator with LSTM to make people happy.
Depression Posts Analysis with NLP
Clustering / Natural Language Processing
Analyzed depression sub-reddit posts with natural language processing. Also made a joke generator in hope to benefit those depressed.