Using Facebook Prophet for Forecasting Natural Gas Production

In this example, we will try to adopt a pure data-based approach to forecast Canada’s natural gas production using Facebook’s Prophet library. The idea of this post is to use a univariate time-series dataset and produce a best-fit model that allows us to confidently predict future production. We will be using the model to forecast the production for the next 10 years i.e. until 2030.

Importing Data and Libraries

We will begin by importing various libraries in Python such as fbprophet, numpy, pandas…