Forecasting Trends

Forecasting trends from data is one of the many objectives data storytellers can identify for their use case. Forecasting involves looking at past trends to identify potential future insights. Let's take a look at how to frame forecasting and how to consequently create the right types of data visualizations for these purposes.

Framing forecasting problem statements

A data storyteller may choose to employ forecasting for several different reasons, including:

• Predicting customer satisfaction for a service.

• Predicting the amount of resources needed for a given product.

Typically, forecasting is visualized as a line plot, with time as the x-axis. However, time does not always need to be the basis for which historical data is compared with the present.

How are forecasting visualizations displayed?

Forecasting visualizations typically involve multiple components:

• Data corresponding to the x-axis, such as a timeline.

• The accurate/historical data.

• The forecasted data using an algorithm, such as ARIMA.

• Upper and lower bounds, such as a prediction interval (optional).

Take a look at a sample forecasting visualization with the Matplotlib library: