As per current design, when scheduled update runs, individual analysis cached with all data preloaded. So even if multiple analysis are there with same IL , they have individual copy of data in it.
Data must be preloaded on Webplayer server (Single Copy) and all information links which uses that data should use the single copy of preloaded data.
There should not be redundant data for such analysis. It is unnecessarily causing increase in disk and RAM
although it's not the exact implementation of this idea, but something like this can be achieved by saving data to the library as an sbdf file and updating that data with automation services: https://community.tibco.com/wiki/caching-data-tibco-spotfire-automation-services