The purpose of this repo is to introduce Facebook Prophet and its application in Business Analytics. Prophet is a Machine Learning algorithm developed at Facebook that enables business analysts to deploy forecasting models on time series data rapidly. Prophet works best on daily time series datasets with seasonal trends.
Why should business analysts be excited about Prophet?
- With an out of the box ability to graph seasonality and trends, Prophet can enable analysts to quickly understand the time-based drivers of their business [e.g. monthly trends]
- Understanding the drivers of KPIs
- Modeling the relationship of different variables and how they affect a KPI [such as # of users driving revenue] can help determine the future actions related to those variables.
- Reliable and Accurate Forecasting
- Remove bias from qualitative forecasting and rely on data-driven modeling
- Automatable and Tunable
- Because the forecasts are generated by an algorithm, little interaction is necessary to produce regularly scheduled forecasts.