The nature of time series data
Why order is information in a time series: serial dependence, one realization of a stochastic process, and why you cannot shuffle the rows.
A 2 minute 38 second animated lesson on what makes time-series data different from a cross section. Built for ECON3006 Economic & Financial Modelling at Western Sydney University as a visual primer and self-reading aid.
The core idea is that in a time series the ordering carries information. The path we observe is one realization of a stochastic process, so reshuffling the rows destroys the autocorrelation structure the whole toolkit relies on. The video walks through temporal dependence (today depends on yesterday and last week), dynamic effects (a shock is spread over future periods rather than landing all at once), and the trend, seasonal, cyclical, and irregular components that can make the mean drift over time.
The takeaway is simple: in time series, the order is the information, which is exactly why the cross-section assumptions need adapting. Pair it with the Atlas concept page for the formal stochastic-process notation, a worked Stata tsset example, and a quick quiz.