Reduce processed dataset size

[gflegar]

To avoid loading unnecessarily large datasets, another JSONata script could be added to preprocess each file before it is loaded. This way, a single huge JSON object is never formed in memory.

It is probably best to implement this issue once #1 is complete.