What is Time Series Forecasting?
Most managers wished they had a crystal ball to look into the future. If they did, their questions may be similar to, “How many patients do we think we will see next month?” or “Which county is expected to have the largest population in the next ten years?” Using historical time series data, forecasts can be made to help predict future volumes.
Most managers wished they had a crystal ball to look into the future. If they did, their questions may be similar to, “How many patients do we think we will see next month?” or “Which county is expected to have the largest population in the next ten years?” Using historical time series data, forecasts can be made to help predict future volumes.
ARIMA Forecasting Demonstration:
To the right is a randomly generated time series data set of patient volumes per week. The objective is to accurately forecast future volumes and possible variation of the projected volumes. Questions to Ask:

Graph Interpretations

Forecasting Graph

Seasonal Plot
<
>
Forecasting Plot
In the forecasting plot, the black line represents historical time series data and the red dashed line represents the forecasted values. The blue shaded areas are the 80th and 95th confidence intervals (i.e., using the 95th confidence interval, you can be 95% confident the forecasted values will fall within this area). 
Seasonal Plot
Given enough data, the model can statistically determine the probability of seasonality. The Xaxis represents months out of the year, the Yaxis represents the volume and each color represents a different year. In the example to the right, you can see a trend of high volumes around July each year. 