profile picture


I'm Christian Normand.

I build impactful data solutions that are
flexible, scalable, and maintainable.

Recent Projects


SmartWatt Home Energy Monitor

SmartWatt is an Arduino-powered energy monitoring solution that delivers to-the-minute data and historical trends through an online dashboard. In addition to designing and building a Wi-Fi enabled measurement device, I set up robust data infrastructure using AWS services like IOT Core, DynamoDB, and Lambda. I used Flask and Plotly Dash for the dashboard application. All the tools I used can be scaled for organizations with massive data requirements, making this an incredibly valuable skill-building project.

This app will be back online soon!
After a successful 6 months of runtime, I'm making some updates and migrating to a new hosting provider.


Image Classification with Active Learning

In this project, I explore active learning as a strategy to reduce the amount of data required to train a convolutional neural network for image classification. This has many real-world applications - specifically for projects where data needs to be collected and annotated. Active learning leads to time savings and cost reduction in the annotation phase, and ultimately ensures you're getting the most value from your data.

View on GitHub

Underrepresentation in Higher Ed

In this analysis project, I explore underrepresentation at California universities through SAT and ACT participation rates. I used data from the California Department of Education to determine which school districts served the largest proportion of students from underrepresented minority groups. I then pulled in exam participation rates to identify 11 school districts that would be a good starting point for organizations working towards equity in higher education.

View on GitHub

Cyberbullying Detection

Can machine learning be used to detect cyberbulling in online comments? I worked as part of a small data science team to answer this question using natural language processing (NLP) techniques. We compared logistic regression, naive Bayes, support vector classifier, and XGBoost models on their accuracy detecting harassment in comments from Wikipedia articles. XGBoost achieved the highest accuracy, averaging 91.6 percent across our three harassment metrics.

View on GitHub

About Me

I love learning how things work. When I was four years old, I would point at every technological contraption and demand to know what it was and how it functioned. My inquisitive spirit has been a driving force in my life ever since.

I studied mechanical engineering in college, but found myself fascinated by data because of its ability to inform, generate insights, and even lead to new discoveries. In 2021, I completed a data science immersive course through General Assembly as a way to build my data science skillset beyond what I'd learned through on-the-job experience and self teaching. The statistical methods and modeling techniques I learned are highly applicable to the projects I will work on throughout my career. Equally valuable, though, is my approach to solving problems, which begins with the same question I asked as a child - "How does it work?"

When I'm not at the computer, you can find me running, cooking delicious food, or playing the drums 🤘.

Get in touch!

Send me a message using the form below, and I'll get back to you shortly.