Mentor: Omar Shehata, Graphics engineer at Cesium

Create a web app to visualize LiDAR data from self driving cars.

LiDAR is a way of obtaining an accurate 3D representation of a scene, often used in self driving cars to detect their surroundings. Lyft recently released a large dataset of sensor readings from their self driving car work: https://level5.lyft.com/dataset/

Most analysis done on data like this is done without visualizing all the data together, because it's in different formats/too large to efficiently visualize in real time. This project would involve taking this data, converting it to a standard format that Cesium's pipeline can ingest (likely using Python since it has many of the helper libraries you'll need), and then building an application to visualize it in 3D using Cesium's JavaScript library.

This would not only help people get more insight out of this particular dataset, but would provide a framework for anyone doing self driving car analysis to see their result and share their research. For inspiration you can see how people are analyzing this data in this Kaggle competition:

https://www.kaggle.com/c/3d-object-detection-for-autonomous-vehicles/discussion/109415

Mentor: Charlie Liu, Student at Yale University

Ever heard of a "falling sand" game? (if not, check this out: https://sandspiel.club) Not only is this a really fun project to work on, but it also incorporates several topics outside of CS, such as physics and basic biology concepts. The project also doesn't have to mimic the traditional "falling sand" style - it could incorporate some twist or new element, such as adding a "person" character or drawing some inspiration from the Sugar, Sugar game (https://www.youtube.com/watch?v=sWCkmqUaZpM). Ready to start building your own sandbox world?

Mentor: Yang Xu, Engineer at LexisNexis Risk Solution

We will design a system to fetch data from REST api and store data in a database. The system is comprehensive with various functionalities such as input/output, caching, logging, etc. The system has to be highly scalable so that it could handle big amount of data.