Prophet model
Facebook’s time series forecasting tool
Introduction
Prophet is a procedure for forecasting time series data based on an additive model where non-linear trends are fit with yearly, weekly, and daily seasonality, plus holiday effects. Documentation
Model
We will use our monthly dataset. First, we shall organise our data, grouping it by date and taking the mean of the average price column and rest the index.
London
Now it is a relatively simple step to create our model. We must specify the data, x, y, frequency and forecasting periods and we’re good to go. Prophet essentially does the hard work for us.
Kingston
We can run a similar analysis on a particular borough. Lets pick Kingston Upon Thames, being of particular interest to myself.
We perform the same steps as we did for our London forecast but now based upon our Kingston data. As a result, we see a similar trend just at higher prices.
