I wouldn't normally worry too much about the minimum as it wouldn't impact your system's capacity but the mean, maximum and the 95th percentile would be important to determine how fast your storage would need to be and also how fast and large your index can grow over time (those are important factors in terms of storage and distribution strategies).
Christian von Wendt-Jensen - 12 years ago
We receive our documents in a steady pace around the clock. It would be helpful to monitor every period compared the running average of corresponding periods in the past, e.g. the last 7 days or the last 14 days. Maybe that is what you mean by "average" in your poll...?
Approximate the document(time) function using the minute by minute data using a supersmoother(supsmu in R). Then report the value of the supersmoother for the midpoint of the time interval for which you want to show an aggregate value.
I wouldn't normally worry too much about the minimum as it wouldn't impact your system's capacity but the mean, maximum and the 95th percentile would be important to determine how fast your storage would need to be and also how fast and large your index can grow over time (those are important factors in terms of storage and distribution strategies).
We receive our documents in a steady pace around the clock. It would be helpful to monitor every period compared the running average of corresponding periods in the past, e.g. the last 7 days or the last 14 days. Maybe that is what you mean by "average" in your poll...?
Approximate the document(time) function using the minute by minute data using a supersmoother(supsmu in R). Then report the value of the supersmoother for the midpoint of the time interval for which you want to show an aggregate value.