Reshape EIA-176 data into tidy (long) format

[davidmudrauskas]

Tidy the table, so that each row represents a single observation of many coincident variables, with one variable per column. The natural primary key identifying each row appears to be (report_year, area, id). Once the table is tidied with clean columns, the line and itemsort information should probably be gone.

[e-belfer]

@davidmudrauskas Happy to chat about the design process here for the transformation, or look over a draft PR. Just nudge me if you need anything!