Background

Based on analysis of empirical data, the Hawaiian Island Chain appears to be moving every year, shifting from its initial position towards Japan. In this worksheet, your goal is to apply concepts from ML to predict some properties of this data; namely:

1. Determine average speed of motion
2. Predict distance from origin at some time

Introducing our Data

Our data is available below as a CSV; it contains 2 columns, which are time(units are in millions of years) and distance, which is the distance travelled from the start point, in kilometers.

data.csv

Hints:

• To load this data, Pandas might be a good tool. pandas.read_csv might be a good place to look…
• Your goal is to make a predictor that predicts position, and also extract some insights about the data. Visualizing the data might help you decide what model to try and fit
  • Scroll down for teaser, or figure out yourself what model makes the most sense

<aside> 💡 Don’t scroll below here if you don’t want the model type spoiled!

</aside>

Fitting our Model

Reccomended Approach:

• Start by solving with SKLearn, and show results to Adi. If answer is correct, you can move one
• Solve with PyTorch. A bit more involved, but gets you in the headspace of gradient descent
• Solve by hand. Needs a bit of calculus, but by this point you know everything that’s going on internally