Add more concrete example in the dataloader notebook

We have had feedback that it would be useful to have a more concrete example in the dataloader notebook about how we would use the dataloader/dataset to make predictions.

We could:
- Add the regular persistence model
- Show a very minimal validation loop using the dataset where we make predictions and calculate MAE scores

Start with the dataloader notebook and review how the dataloader and dataset are currently introduced. Add a concrete prediction example using the regular persistence model and a minimal validation loop, then confirm that it produces predictions and reports MAE scores.

Written by the indexing model from the issue text.