Body
Body
Document
Body

How to Build a Strong Data Science Portfolio in the Rice MDS Capstone Experience

See how Rice MDS Online grad students are exercising applied data science/ML skills and building strong portfolios in the hands-on Capstone course.

MDS Blog Hero Capstone

Rice U. data science master’s students create ML model to predict Airbnb rates in Austin Days before graduating as the first Rice University alumni in the Master of Data Science Online program, three students presented their final capstone project – predicting Airbnb prices in Austin-- and published their findings as a machine learning research paper.

“The pricing of Airbnb rentals is prone to high fluctuations, with prices changing frequently based on demand, seasonality, and other factors,” said Seifey Mohammad. “Our goal for this project was to predict the prices of Airbnb rentals using a machine learning modeling approach.”

The MDS trio began by digging into a deeper understanding of what drives price in a dynamic market and determined listings, calendar, and reviews would be their focus.

“We chose a sentiment analysis tool to identify the most popular features,” said Kimberly Villegas. “Using three scores – positive, negative, and neutral – and a Bayes dictionary for sentiment reasoning helped us visualize the most important customer preferences based on previous reviews.”

The team’s final ML model utilized a random forest algorithm which allowed them to combine multiple outputs into a single decision - the price point in their model. Incorporating cross-validation tuning reduced runtime costs, although it slightly decreased the accuracy as a result.

Students explore public Airbnb data and search for insights

“Like all data scientists, we faced challenges with data availability and access. Insideairbnb.com installed a paywall in front of historical data prior to the year 2023,” said Sam Chapman. “The restricted access to their previous data limited its use with regard to calendar pricing information, so our final datasets were sourced from insideairbnb.com, GitHub, and Kaggle.”

He said determining which features made a significant difference in the pricing to be both interesting and surprising. “Understandably, locational features made the biggest impact, but the day of the week and time of the year significantly affected the pricing as well.” Managing the data was the most interesting aspect of the problem for Villegas. She said, “In the Airbnb project, I discovered that I enjoyed the process of cleaning and refining the data the most. It's like bringing order to chaos, making the data ready for analysis. This step is crucial because it lays the foundation for accurate insights and predictions.”

Having shared courses, faculty members, mentors, and resources like Slack and Discord channels with in-person MDS Rice students, the MDS Online students joined their peers for a hybrid capstone showcase, recording their presentation ahead of time to avoid connection issues. Their video kicked off the evening’s presentations.

Capstone director celebrates program’s first hybrid showcase

Arko Barman, the director for Rice’s Master of Data Science Capstone Program, expressed enthusiasm for the first event to include teams from both the in-person and online courses. The hybrid environment also allowed local team members with schedule conflicts to participate via Zoom.

omds capstone

“This was the first time that we were offering the capstone course for online students. Offering a course like this with several components is a big challenge for both the instructors and the students. I want to commend the online students on having completed this course successfully and finishing a large scale data science project collaboratively,” said Barman.

“The online cohort of students proposed their data science project and defined their goals and objectives using open source data with the help of the instructors. This team of students then designed a project plan and executed it successfully to finish the project using the advice and guidance from Dr. Xinjie Lan and me.”

Making the most of the Master of Data Science online course options

Chapman said even before the capstone project, he had already begun to see himself as a data scientist. “The first time it hit me that I was actually ‘doing data science’ was when I was enrolled in both Big Data and Machine Learning at the same time. Working with AWS and the complicated neural networks --that is when it really felt real to me.”

For Mohammad, his biggest ‘ah-hah’ moments occurred during the capstone project. He said, “Our team had to troubleshoot problems and try new approaches to programming we may not have attempted before. But when our approaches worked, it was the best feeling in the world. That course honed my ability to be adaptable and reinforced my confidence in self-learning.”

Villegas tailored her program around interests she discovered during a Google data science internship. “The Google internship I completed before my MDS program at Rice was a pivotal experience,” she said. “It exposed me to the real-world applications of data science and opened my eyes to the complex issue of algorithmic bias and its implications.

“This experience shaped my course selections, pushing me to delve deeper into areas like fairness and ethics during my studies. I also took an ethics course as an elective to gain a better understanding of the ethical considerations in engineering as a whole and its impact on society."

Domain knowledge impacts data science approach

Before shifting her focus to data science, Villegas had spent six years as a radiologic technician. She said her experience in the medical technology and imaging field definitely left a mark on how she approaches data science.

“Working as a radiologic technician gave me this deep appreciation for accurate data and how it applies in the real world. It's like behind-the-scenes knowledge that has shaped how I view and address data science tasks. I'm all about solutions that don't just stay on paper, but actually matter in practical situations. This background has enriched my analytical thinking and strengthened my dedication to using data responsibly and ethically. It's become second nature in every data science journey I take.”

Mohammad was working as a data analyst in the finance industry when he applied to the Rice OMDS program. In addition to leveraging his domain knowledge with additional data science skills, he believed obtaining an advanced degree and broader intellectual horizons would support and speed up his career progression plans. “Rice’s OMDS program was a way for me to get the data science skills I wanted in a professional setting and at an accelerated pace,” he said.

Shifting from startups to data science

The local startup scene influenced Chapman’s decision to pursue a data science graduate degree. He said, “After spending time working with a group of serial entrepreneurs across Houston, I knew there was more that could be done with the data I was using. I asked around and one of my mentors, Alec Walker, introduced me to the world of data science. Alec mentioned careers in data science and recommended Rice, so I researched this and other programs.

“Rice’s impressive STEM programs made my decision easy. Now that I’m searching for my first data science job, I can look back at the program and see how fantastic it was. We learned tools and techniques from a healthy mixture of data science researchers and professionals with extensive experience in the field. I feel the future is bright with an MDS from Rice, and I am excited for what is to come.”

Working full-time while pursuing a master’s degree

Having worked full-time while pursuing his master’s degree as a part-time student, Mohammad advises current and prospective students to “keep grinding. It can be tough to do such a rigorous program--especially while working full time--but there is a light at the end of the tunnel. You will have to hone your time management skills. Be ready for long nights and busy weekends, but it will be worth it.”

Rice University is proud to offer one of the only top-ranked 100% online Master’s in Data Science programs with access to real-world data sets, small class sizes and 1:1 interaction between grad students and faculty. To learn more about the Rice MDS Online student experience, curriculum, admissions and more, visit the Rice MDS Online website or fill out our request form to connect with a coach.

Build your data science skills & portfolio in Rice's hands-on OMDS program.
×
Body

Request Information

Send us your information and a Rice Enrollment Coach will follow up with you.

Loading...