Hacking Through Asteroids

Our Logo
Dashboard

What it does

Hacking Through Asteroids is a web application that visualizes our data analysis of EOG's asteroid mining dataset. Specifically, it shows how economically each asteroid was mined and gives drill bit efficiencies in terms of feet/dollar and feet/hour. Because how economic the mining of an asteroid is is based on both the cost and time taken to mine the asteroid, we give users the ability to choose how to prioritize the two in our evaluation of economic viability. The app also uses regression models to help predict how to change hook load, differential pressure, and weight on bit to best improve rate of penetration of drill bits using real-time data.

How we built it

The data analysis behind Hacking Through Asteroids was done using several data science modules in the Python programming language. We used them to calculate the cost and time associated with each asteroid drilling as well as the drill efficiencies described above. We then experimented with a number of prediction approaches, including correlation plots, linear regression, polynomial regression, recurrent neural networks with long short-term memory, and more. For the front-end, we used Angular.js as our main framework, nebular as our UI library, and e-charts for graphs to show our data analysis. We also utilized FastAPI to connect our backend to our frontend.

Challenges we ran into

Most of the challenges that we ran into concerned the nature of the dataset. We found no strong correlations between variables, which made it difficult to make clear predictions about how changing variables would affect rate of penetration. In addition, the dataset was both very large and noisy, making it difficult to create predictions. Furthermore, our dataset does not include any variable that describes the degradation of a particular drill bit besides rate of penetration, which is already dependent on the other variables (which can be changed independently). Collectively, these challenges made getting our rate of penetration predictions more difficult than expected.