The concept of a locustree stems from the fact that it does not make sense for us to represent the full (remember 1D) dataset on a screen, it is simply too small for that. And even if we would draw all the points, we would not be able to notice any remarkable aspects of our data. In other words, it makes sense for us to look at the data at different zoom levels. At the highest zoom level we see individual data points, the lowest zoom level gives an aggregate idea of the full dataset and different zoom levels can exist in-between.
For more information, see this blogpost