Statistics and Data Science I (ABC) CHEAT SHEET

https://bit.ly/r-cheatsheet-abc
Word Equations Summary Tables Simple Statistics
outcome = explanatory + other stuff # compute five-number summary mean(data_set$Y)
Y = X + other stuff favstats(~ Y, data = data_set) var(data_set$Y)
sd(data_set$Y)
# create frequency table
Basics tally(data_set$Y) cohensD(Y ~ X, data = data_set)
tally(~ Y, data = data_set) cor(Y ~ X, data = data_set)
print("Hello world!")
# tally by condition b1(Y ~ X, data = data_set)
# assign value to object b1(one_model)
tally(~ Y < 1900, data =
myNumber <- 5
data_set)
pre(Y ~ X, data = data_set)
# combine values into vector f(Y ~ X, data = data_set)
# two-way frequency table
myVector <- c(1, 2, 3)
tally(Y ~ X, data = data_set, margin =
TRUE, format = “proportion”)
# first element in vector
myVector[1]
Data Frame # arrange rows by variable
arrange(data_set, Y)
# orders values or cases # structure of data frame # arrange rows by variable in descending
sort(myVector) str(data_set) arrange(data_set, desc(Y))
# arithmetic operations
# view first/last six rows # creates data frame from csv file
sum(1, 2, 100), +, -, *, /
head(data_set) data_set <- read.csv("file_name", header =
sqrt(157)
abs(data_set$Y) tail(data_set) TRUE)

# logical operations # select multiple variables # convert quantitative variable

>, <, >=, <=, ==, !=, |, & select(data_set, Y1, Y2) # to categorical
factor(data_set$Y)
# results in a variable with values # first six rows of selected variables factor(data_set$Y, levels = c(1,2), labels
# of TRUE or FALSE head(select(data_set, Y1, Y2)) = c("A", "B"))
data_set$C <- data_set$A > data_set$B
# transform values
# select variable (a column) recode(data_set$Y, "0" = 0, "1" = 50, "2" =
data_set$Y 100)

# find rows that meet condition # creates two equal sized groups
data_set[data_set$Y > 40] ntile(data_set$Y, 2)
filter(data_set, Y > 300)
filter(data_set, Y != "NA") # convert categorical variable
# to quantitative
as.numeric(data_set$Y)
 Probability Distribution
# CI using t distribution
# calculate the probability area confint(empty_model)
xpnorm(65.1, data_set$mean, data_set$sd)

# calculate a z-score # calculate p-value using F-distribution

zscore(data_set$Y) xpf(sample_F, df1 = 2 , df2 = 10)

# returns t at this probability

qt(.975, df = 999)
# returns F at this probability
qf(.95, df1 = 1, df2 = 100)

Page: 2 ▷ Updated: 2023-04 ▷ Learn more about CourseKata @ https://coursekata.org

 Simulation Fitting and Evaluating Models
# randomize sampling distribution
# sample without replacement # of b1s, centered on 0 # empty model
sample(data_set, 6) sdob1 <- do(1000) * empty_model <- lm(Y ~ NULL,
b1(shuffle(Y) ~ X, data = data_set) data = data_set)
# sample with replacement
resample(data_set, 10) # bootstrap sampling distribution of # use one expanatory variable
b1s, one_model <- lm(Y ~ X, data = data_set)
do(3) * resample (data_set, 10) # centered on sample b1
sdob1_boot <- do(1000) * # create a function from a formula
# mixes up values in a variable b1(Y ~ X, data = resample(data_set)) one_model_fun <- makeFun(one_model)
shuffle(data_set$Y)
one_model_fun(x_level_1)
# count the number of b1s at the upper
# simulate sampling 10000 Ys # and lower extreme
# model predictions and residuals
# from normal distribution tally(sdob1$b1 > sample_b1 | data_set$empty_predict <-
sim_Y <- rnorm(10000, Y_stats$mean, sdob1$b1 < -sample_b1) predict(empty_model)
Y_stats$sd) data_set$empty_resid <-
resid(empty_model)
# return TRUE for middle 95% of
# put simulated Ys into dataframe
data_set<- data.frame(sim_Y) distribution
# produce ANOVA table
middle(sdob1$b1, .95)
anova(empty_model)
# simulate sampling distribution of supernova(one_model)
means # randomize sampling distribution of
sim_SDoM <- do(10000) * mean(rnorm(157, PREs # t-test, using pooled variance
Y_stats$mean, Y_stats$sd)) sdoPRE <- do(1000) * PRE(shuffle(Y) ~ X,
data = data_set) t.test(Tip ~ Condition, data =
data_set, var.equal=TRUE)
# bootstrap sampling distribution of
# randomize sampling distribution of
means # pairwise comparison
bootSDoM <- do(10000) * Fs
sdoF <- do(1000) * # corrections: "Bonferroni",
mean(resample(data_set$Y, 157)) “Tukey” (default) or "none"
fVal(shuffle(Y) ~ X, data = data_set)
pairwise(one_model, correction =
# counts extreme Fs "none")
tally(~fVal > sample_F, data = sdoF)

Page: 3 ▷ Updated: 2023-04 ▷ Learn more about CourseKata @ https://coursekata.org

 Visualizations I
# faceted grid of histograms # density histogram
# histogram gf_histogram(~ Y, data = data_set) %>% gf_dhistogram(~ Y, data = data_set, fill
gf_histogram(~ Y, data = data_set) %>% gf_facet_grid(X ~ .) = "orange") %>%
# change labels gf_density()
gf_labs(title = "Graph Title", x
= "Y_Name", y = "Frequency")

# bar graph # boxplots # point plot

gf_bar( ~ Y, data = data_set) gf_boxplot(Y ~ X, data = data_set) gf_point(Y ~ X, data = data_set)

Page: 4 ▷ Updated: 2023-04 ▷ Learn more about CourseKata @ https://coursekata.org

 Visualizations II
# boxplot overlaid with jitter plot # sampling distribution of b1
# jitter plot gf_boxplot(Y ~ X, data = data_set, fill gf_histogram(~b1, data = sdob1, fill =
gf_jitter(Y ~ X, data = data_set) = "orange") %>% ~middle(b1, .95)) %>%
gf_jitter(height = 0, alpha = .2, size # modify the limits on x- and y-axes
= 3) gf_lims(x = c(-12, 12), y = c(0, 70))

# Add model # CI for pairwise comparisons

gf_point(Y ~ X, data = data_set) %>% pairwise(one_model, plot = TRUE)
# add model predictions as red points
gf_point(Y ~ X , shape = 1, size = 3,
color = "firebrick") %>%
# add best fitting model as a red line
gf_model(one_model, color = “red”)

Page: 5 ▷ Updated: 2023-04 ▷ Learn more about CourseKata @ https://coursekata.org