Notes
The estimation date for actual data is set two years ahead of the observation year. This discrepancy arises because official data is typically published with a one-year lag, and the data is presented as three-year centered moving averages. Despite updates to this data, the values should ideally remain consistent. However, it's worth noting that the Prevalence data is directly downloaded from FAOSTAT after release of the SOFI report, and our team has observed that figures appearing later in the World Development Indicators sometimes deviate from this data and backward corrections may have been applied, resulting in changes to the actuals in our dataset across different versions. Similarly, the population counts can be backward revised when the United Nations releases a backward revision of national population counts. For the most up-to-date official data, users should always refer to FAO as their primary source. Historical estimates are consistently updated in each version of this dataset, as the model used to produce these estimates undergoes retraining with new data. Additionally, data derived from IPC-compatible sources is also subject to updates, as the linear regression model used for this data is re-estimated whenever new data becomes available.
