Data Summarization

Suppose we use a trust rating with an integer in $[1,10]$ to represent the reputation of an eBay seller on a day. With such data, we can draw a spot in a two dimensional space, where the x$-$axis is the date and the y$-$axis presents trust value. We also suppose there is a large dataset of trust ratings for each seller for a period of $5$ years.

If a popular seller receives a large number of trust ratings each day, introduce a data summarization approach which aggregates these ratings so as to reduce the size of the trust rating dataset while the seller$$s trust degree and reputation in a certain period $($say $5$ years$)$ can be well represented.