Along with reduced file sizes and storing as pickle files instead of .csv, I think opening from file directly in the callbacks that are creating content for each tab is going to be faster than storing in a hidden div. The hidden div stores data on the clients browser, so there is a lot of network back and forth to retrieve these datasets. The pre-aggregated pickled data frames should be small enough to support fast loading.
Along with reduced file sizes and storing as pickle files instead of .csv, I think opening from file directly in the callbacks that are creating content for each tab is going to be faster than storing in a hidden div. The hidden div stores data on the clients browser, so there is a lot of network back and forth to retrieve these datasets. The pre-aggregated pickled data frames should be small enough to support fast loading.