How to predict future values with linear regression in R

Question

Writing my first R program and I'm stuck. Need to forecast the population for 2018 through 2022 using linear regression. Getting errors trying to use predict().

Here's what I have so far:

X <- c(2010, 2011, 2012, 2013, 2014, 2015, 2016, 2017)
Y <- c(11539282, 11543332, 11546969, 11567845, 11593741, 11606027, 11622554, 11658609)
model.1 <- lm(Y ~ X)
summary(model.1)

plot(X, Y, ylim=c(10000000,13000000))

lines(sort(X), fitted(model.1)[order(X)])

Answer 1

# create a data frame to store your variables
df <- data.frame(
  X = 2010:2022,
  Y = c(11539282, 11543332, 11546969, 11567845, 11593741, 11606027, 11622554, 11658609, rep(NA, 5))
)

# check the data frame
df
#   X        Y
# 1  2010 11539282
# 2  2011 11543332
# 3  2012 11546969
# 4  2013 11567845
# 5  2014 11593741
# 6  2015 11606027
# 7  2016 11622554
# 8  2017 11658609
# 9  2018       NA
# 10 2019       NA
# 11 2020       NA
# 12 2021       NA
# 13 2022       NA

# The lm function in R will exclude the observations with NA values while fitting the model
model.1 <- lm(formula = Y ~ X, data = df)

# get the model summary
summary(model.1)

# broom is an extremely useful package for handling models in R
# install.packages("broom")
# tidy your model and include 95% confidence intervals
broom::tidy(model.1, conf.int = T)
#          term     estimate   std.error statistic      p.value     conf.low    conf.high
# 1 (Intercept) -22799768.60 3272284.123 -6.967539 0.0004342937 -30806759.40 -14792777.80
# 2           X     17077.01    1625.171 10.507824 0.0000436377     13100.36     21053.66

# The model is of the form: Y = - 22799768.60 + 17077.01 * X
# you can get rough predictions for 2018 through 2022 using this formula:
- 22799768.60 + 17077.01 * 2018:2022
# [1] 11661638 11678715 11695792 11712869 11729946

# you can use the predict function as well for precise predictions
# get predictions for every X value
predict(object = model.1, newdata = df)
#        1        2        3        4        5        6        7        8        9       10       11       12       13 
# 11525025 11542102 11559179 11576256 11593333 11610410 11627487 11644564 11661641 11678718 11695795 11712872 11729949

# get predictions for 2018 through 2022 
predict(object = model.1, newdata = subset(df, X >= 2018))
#        9       10       11       12       13 
# 11661641 11678718 11695795 11712872 11729949