I'm involved in research on a demolished cemetery in Baltimore. We have been transcribing death certificates to understand the burial population. The problem is that death certificates only started to be recorded 23 years after the cemetery was established. We have 30,000 records of burials after that point. I plotted the number of burials per quarter from the records we had transcribed and in minitab, I did a trend analysis fitting different regression models, and the data best fits an exponential decay in rates of burials. The question is, how can I estimate what the curve would look like for the first 23 years for which we have no data. Is there a way to use the shape of the curve of the best fit line I have, together with a known 0 for the quarter before burials began to estimate how many burials may have taken place in the first 23 years? I feel it should be possible but I've approached the limit of my math/stats knowledge.…