In my recent foray into the world of time series analysis, I took a deep dive into the patterns of police shootings across the United States. The first step in this journey was to determine if the data at hand was actually predictable – for this, I turned to the Dickey-Fuller test. The results were nothing short of a green light:
Test Statistic -20.379884
p-value 0.000000
With a test statistic significantly below the critical value and a p-value of zero, I confidently rejected the null hypothesis. What this means is that the data was stationary, indicating that the number of police shootings did not depend on time, making my dataset ripe for further analysis and, hopefully, for developing forecasts that could hold up in the real world.
Be First to Comment