After many, many, many hours of deliberation, we were inspired by our mentor on a goal to achieve, and we contemplated solutions, coming to a result of a simple health algorithm to sort devices for an app.
What it does
It takes large data ~250,000 lines and groups the data, by the device name, and then divides the device's stats into smaller tables for more acute and precautionary data analysis.
How I built it
It was build with MATLAB, further, it had originally started out as CPU usage vs timestamp, as we had already implemented the smart sort of device names. We then took a break, and came back with our new found inspiration to build upon this to create an application which can look in depth at a single device and its trends or an entire app and its devices.
Challenges I ran into
The goal. What were we supposed to do? How to rank device health. How were we supposed to predict the data with the limited amount of information? Further, we had misaligned elements such as the end date and timestamp.
Accomplishments that I'm proud of
We successfully created a super fast algorithm to rank device health over hundred thousands lines of code, using smart iteration and using MATLAB's built in analysis tools.
What I learned
We learned about network pathing and load balancing. As well, we learned how to separate large data into smaller subsets to analyze them just as powerfully but much quicker. We also learned new techniques for indexing and seeking information via MATLAB tables.