r - 在 R 中为时间序列的变化创建指标变量的最简单方法

Question

我有一个包含 1400 万行产品、关税税率、贸易量和年月组合的数据集，格式如下：

df <- as.data.frame(matrix(c(1220, "2013-1", 10011900, 29307, .1,
                   1220, "2013-2", 10011900, 28202, .1,
                   1220, "2013-3", 10011900, 22383, .15,
                   1220, "2013-4", 10011900, 21303, .15,
                   1220, "2013-5", 10011900, 21201, .15,
                   1220, "2013-1", 10019900, 9960, .12,
                   1220, "2013-2", 10019900, 10043, .12,
                   1220, "2013-3", 10019900, 11001, .1,
                   1220, "2013-4", 10019900, 10997, .1,
                   1220, "2013-5", 10019900, 12038, .1), 
                 ncol = 5, byrow = T))
colnames(df) <- c("country", "date", "product", "value", "rate" )

我正在尝试在数据中添加一列，这样我就可以用来创建一组指标变量，标记关税税率发生变化之前/之后的几个月。所以，上面看起来像这样：

df_transformed <- as.dataframe(matrix(c(1220, "2013-1", 10011900, 29307, .1, -2, 
                                        1220, "2013-2", 10011900, 28202, .1, -1,
                                        1220, "2013-3", 10011900, 22383, .15, 0, 
                                        1220, "2013-4", 10011900, 21303, .15, 1, 
                                        1220, "2013-5", 10011900, 21201, .15, 2,
                                        1220, "2013-1", 10019900, 9960, .12, -2,
                                        1220, "2013-2", 10019900, 10043, .12, -1,
                                        1220, "2013-3", 10019900, 11001, .1, 0,
                                        1220, "2013-4", 10019900, 10997, .1, 1,
                                        1220, "2013-5", 10019900, 12038, .1, 2)))
colnames(df_transformed) <- c("country", "date", "product", "value", "rate", "months_since_change")

我不确定如何最好地找到关税变量何时发生变化并基于此创建一个新列。

谢谢您的帮助！

标签： rtime-seriespanel-data