Some data series are very volatile. That is, they don’t follow a smooth or step-by-step pattern. And it’s difficult to draw conclusions when new data are added to a volatile series. The weekly release of initial claims for unemployment insurance is a great example. In this and similar cases, it is useful to adopt some kind of smoothing mechanism: Here we provide a four-week moving average. Traditionally, a moving average is centered—say, the average of two periods before and two periods after. This moving average takes the last four observations, which allows you to better read trends, especially if you’re focusing on the most recent data. Of course, trends become more obvious if you look at longer spans of time. This graph shows a span of five years. Narrow or expand the sample with the slide bar to see how a moving average can help you interpret the data and avoid the pitfalls of volatility.

How this graph was created: Search for “initial claims,” select the two (seasonally adjusted) series, and add them to the graph. Finally, restrict the sample to the last 5 years, which is done by using the settings above the graph on the right.

Suggested by Christian Zimmermann