iNZight Overview
Remember we are showing evidence of each component of the statistical cycle.
Focus on the discussion and interpretation in context.
(1) Input some data
Discussion: Of the Long Term Trend in very general terms but in context
Discuss relative effect of seasonal effect vs long term trend
Residuals - Any unusual observations - do they warrant further investigation
Discuss the estimated seasonal effects ie what one seasonal cycle is like and possible reason why? Discussion in context
Discuss the individual seasonal effects and how they may have changed over time.
Discuss the predictions in context
Discuss the accuracy or margin or error of the predictions
Discuss how might the forecasts be used (and who might use them)
Discuss what you notice about the comparative data series - similarities, differences, possible relationships, reasons, causes, links etc
Discuss and compare the Trend Lines.
Discuss and compare the Average Seasonal Effects (red line) and the Seasonal Effects for each cycle (gray lines)
Discuss why the new variable has been added.
Discuss the further insight and information provided by the new variable
Discuss whether the actual data values are within the 95% confidence intervals for the new predictions - or not.
Are the estimates above or below expected?
(2) Make a graph the time series data with trend line. (find the gradient)
(3) Decompose the data into trend, seasonal & residual. (work out the % contribution of each component).
(4) Graph the Individual & Estimated Seasonal effects.
(5) Make Predictions of the next two cycles of data (with confidence intervals)
(6) 'Compare Series' between different variables that are available.
(7) Combine variables to make a new variable into the series and analyse.
(8) Test model Robustness by removing the last three data values then estimating them again