DataNew in edition 1.1

`balance_sensitivity.csv`

Tornado table from the QNT model: change in 2050 soybean-meal and maize import need when one assumption (or a paired set) is varied, all else at S-BASE, plus an S-ALT comparison row. Regenerate with tools/balance_model.py.

13 rows · 20 columnsDownload CSVUsed in: Protein balance 2050, W. Balance model

Columns

Column	Meaning
`record_id`	Stable row ID assigned by the consolidation step (prefix BLS plus row number in the stream's row order).
`rank`	Rank by swing in 2050 soybean-meal import need (1 is the largest); "n/a (comparison)" for the S-ALT row.
`assumption_varied`	Assumption or paired set varied.
`scenario`	Scenario the run starts from (S-BASE, or S-ALT for the comparison row).
`low_case`	Parameter values in the low run (SENS-LOW rows of balance_assumptions.csv).
`high_case`	Parameter values in the high run (SENS-HIGH rows of balance_assumptions.csv).
`sbm_import_2050_low_mt`	2050 soybean-meal import need in the low run, Mt.
`sbm_import_2050_central_mt`	2050 soybean-meal import need in the central scenario, Mt.
`sbm_import_2050_high_mt`	2050 soybean-meal import need in the high run, Mt.
`sbm_swing_mt`	Absolute difference between the high and low runs, Mt soybean meal.
`maize_import_2050_low_mt`	2050 maize import need in the low run, Mt.
`maize_import_2050_central_mt`	2050 maize import need in the central scenario, Mt.
`maize_import_2050_high_mt`	2050 maize import need in the high run, Mt.
`maize_swing_mt`	Absolute difference between the high and low runs, Mt maize.
`horizon_year`	Year or years the value or statement refers to.
`foresight_type`	Foresight type from the futures protocol: trend, projection, estimate, signal, wildcard or vision; "projection (official target)" marks a government target.
`evidence_label`	VN-direct (evidence about Vietnam), VN-adjacent (comparable country, transfer assumed) or general. Some rows give a label per part of the row.
`confidence`	High (primary source read), Medium (reputable secondary or primary read in part) or Low (single claim, press, company marketing, or derived by us from weak inputs).
`source_ids`	Source IDs (see sources.csv) supporting the row, separated by semicolons. Free text such as "QNT model" or "our calculation" marks our own work.
`notes`	Free-text notes from the research agent.

Filter rows

The interactive table needs JavaScript. You can still download the CSV. Download CSV

record_id	rank	assumption_varied	scenario	low_case	high_case	sbm_import_2050_low_mt	sbm_import_2050_central_mt	sbm_import_2050_high_mt	sbm_swing_mt	maize_import_2050_low_mt	maize_import_2050_central_mt	maize_import_2050_high_mt	maize_swing_mt	horizon_year	foresight_type	evidence_label	confidence	source_ids	notes