Three Awesome Projects from Udacity’s Data Scientist program!

I’m very pleased to share amazing projects completed by students from our Data Scientist Nanodegree program: Ayush Subedi, Payton Soicher, and Genevieve Hayes.

Udacity students are constantly impressing us with the skill and ingenuity they demonstrate when they complete the projects from their Nanodegree programs. And they aren’t just impressing us! These hands-on projects — an integral part of every Nanodegree program — are also an incredibly important way in which our graduates impress recruiters and land dream roles in transformational industries.

Check out these projects from students in the Data Scientist Nanodegree program at Udacity. You’ll hear from the students about why they chose to tackle their project brief in a particular way, and you’ll hear from our Udacity experts, on what makes the students’ work so special.

It’s a Bird.. It’s a Plane.. It’s Superman!

By Ayush Subedi

The image classification project in the Data Scientist Nanodegree program challenges students to train a deep learning model to classify images of flowers. When Ayush tackled this challenge, he decided to take what he learned one step further—testing out his new model on pictures of birds, planes… and Superman! Ayush explains that he took this approach to push his skills further, and to test what he had learned:

“I have tried to make the project very modular, validated inputs, and basically gone the extra mile. The project was very challenging.”

Udacity Content Developer, Juno Lee, was particularly impressed that Ayush extended what he learned to a new dataset, demonstrating a solid understanding of PyTorch, and infusing a lot of fun into his project.

“I love seeing students applying what they learned to new situations. The ability to extend new skills to novel situations is what will make you successful in your career.”

The mathematician’s guide to arguing with an umpire

By Payton Soicher

In the Data Science blog post project, students learn that communicating your findings is just as important as discovering useful data insights. In this project, students are tasked with writing a blog post demonstrating their findings in a clear, informative, and entertaining way.

Many students choose to use one of the data sets provided by AirBNB (which produces one of the most well-known data blogs), but Payton went above and beyond to build a web scraping program to create an original dataset on college softball and baseball stats.

Payton describes this challenge:

“This project is taking Division 2 college baseball and softball play by play results and breaking them out into more detailed data to look for pitch sequences, ball and strike counts, results, and result locations.”
“This was important to me because I played Division 2 baseball, and our scouting reports were always horrible. We could only go off top level aggregates, but that doesn’t give the whole view of a hitter. So I wrote my own web scraping program that would just take in the URL of a college baseball game and return a dataframe of each play by play and other characteristics using some text analysis.”

Data Science Curriculum Lead, Josh Bernhard, praised Payton’s efforts to build a custom web scraper for the project, and points out what a valuable skill it is to develop:

“Building a custom web scraper is impressive; a lot of startups end up doing their own scraping, so it’s definitely a skill you might need in your career!”

Classifying Disaster Response Messages

By Genevieve Hayes

In term two of the Data Scientist Nanodegree program, students dive into Data Engineering and building ETL pipelines. In this project, students apply these skills to build a machine learning pipeline that classifies messages sent during a natural disaster. The goal is to direct each message to the relief agency that can provide the quickest assistance. This project was created in partnership with Figure Eight, and uses real message data from real situations.

Genevieve describes her own approach to the project here:

“My approach to the project was to just take things one step at a time. There were three main parts to the project: building the ETL pipeline; building the ML pipeline; and then building the final app. Thinking about the entire project as a whole made it all seem overwhelming, so I only worried about each part as I came to it. This made everything much more manageable.”

Speaking about the project, Juno Lee says Genevieve’s structured approach means she did a stellar job bringing together all the different components of this project, with solid analysis, machine learning pipelines, documentation, and visuals.

“I’m proud to see such thorough data science work and development practices applied to this meaningful topic. I would love to see this quality of work continue to appear in good causes like this!”

As you can see, the Data Scientist Nanodegree program students are producing some fantastic work. Well done Ayush, Genevieve, and Payton! Learn more about the Data Scientist Nanodegree program.