When I was young there were no special packages to design software. If I had to solve equation or analyze statistics, I searched in the books to find algorithm. Even application was complicated, it usually did not require following support. The most complicated application was for correlation analysis. I supported it about 5 years, not like programmer, but like expert in the statistics.
Something changed when we started to use databases. Of course, database keeps data that changes all time, and this requires procedures to preserve these data. However there is something else. When we start database design we see specific structure of data. During of data accumulation we may see some pattern that was hidden from us initially. And this is task of programmer to help user to discover this pattern.
Assume you work in hospital and your task is put data about patient into database. Monthly you create report which includes charts of changes blood pressure and heartbeat. From chart you can see correlation between blood pressure and heartbeat. Scince that time you together with user begin to study if this correlation is strong, does it depend on patient or disorder, is it possible to use this correlation for forecast. Certainly, this research is impossible until you do not accumulate data in the database.