Well, we know entanglement happens because of experiments. In those experiments, researchers had to somehow independently verify that the entangled states produced predictable results, thereby demonstrating the entanglement.

If you set up the first part of experiment, wherein you know certain particles are entangled, would it be possible for someone on one side of that entanglement to read a pattern from that side alone (if they 'knew where to look') which matched a pattern on the other side?

At this point, if my question even makes sense, we're only talking about a very impractical method of data storage... I'm just curious whether even that is possible.