Statistics data will be read in from a CSV to be provided by departments
and then converted into a chart in the statistics block on landing
pages.

It is assumed that users will not need to be able to download the CSV
files directly and that a link will be provided to an externally hosted
data source, e.g. ONS data.

It's unclear exactly how the data from the CSV files will be used to
display the graphs, so for the moment the unmodified rows are just being
returned.

The block should render twice as two statistics have been added to the
sample content item data.

A new "rows" method will be created in the next commit

Divides the rows method to a small public rows method and a private
`row_lines` method. This allows further refactoring without having to
change the tests.

Use a map rather than an each in rows method to make the method
slightly smaller. `each_with_index` has been replaced with
`each_with_object` for the same reason.

The CSV rows are being returned with symbol keys to that `to_sym`
doesn't need to be used in multiple places.

Finally, the row keys and values are being obfuscated.

We can't guarantee what the keys will be named so it's better not to
rely on them.

We're still assuming that each row will have 3 columns:

- x-axis value
- variable / line name
- y-axis value