Figuring out when Year-To-Date is predictive of overall trends.
I'd love to know the following, which I think might help the issue:
1) What is the r^2 of YTD data each month predicting final YTD data (the aggregate not city by city)?
2) What is the r^2 of each month at predicting final YoY changes? I.e. if murders are down in June 2022 vs June 2021 is that more predictive compared to murders down in August 2022 alone vs. August 2021 alone? The volume of murders is obviously higher in the summer so I am wondering is are we just gaining by getting closer to the end of the year and that's the cause of the predictive gain or are we gaining because early in the year the volume is lower and so just not as informative.