R Data Understanding

Data Understanding (1-6)
#1. Industry data belongs to; 2. Observations and variables count: ; 3. Data types 
#4. Levels: ; 5. Any negatives (when ecommerce data); 6 Find Nulls

#1 Industry data belongs to:
#2 Observations & Variable Count
```{r}
library(skimr); skim(df99)
library(psych); describe(df99)
summary(df99)
```
#3. Find datatypes 
```{r}
glimpse(df99)
df99 %>%
  dplyr::select(where(is.factor)) %>%
glimpse()

find_datatype = function(df){
  sapply(df, FUN=function(col) class(col))
}
find_datatype(df99)
```
#4.a
```{r}
find_keys <- function(df){
  # get count of distincts per each column
  df_distincts <- sapply(df, n_distinct)
  # make tibble with of above distincts and colnames
  full_count <- data.frame(colnames(df), tibble(df_distincts))
  colnames(full_count) <- c('field','cnt')
  # is a column's unique count the same as the df row count
  full_count$all_unique <- (full_count$cnt == nrow(df)) 
  full_count <- full_count[,c(1,3,2)]
  
  return(full_count)
}
df_uniques <- find_keys(df99)
```
#4.b Find Non-numeric Uniques 
```{r}
find_unique = function(df){
  sapply(df, FUN=function(col) if(class(col)=="character"||class(col)=="integer") sort(unique(col)) else{class(col)})
}
find_unique(df99)
```
#4.c Find All Uniques & Coerce to factor 
```{r}
#4.c.1 Selected variables
find_all_unique = function(df){
  sapply(df, FUN=function(col) sort(unique(col)) )
}
#get names of df
names(df99)
#build a vector for re-use
cols_in_question <- c("Survived","Sex","SibSp","Embarked")
find_unique(df99[cols_in_question])
#covert to factors
df99[cols_in_question] <- lapply(df99[cols_in_question], factor)
#4.c.2 Convert all characters to factors [optional]
mutate_if(is.character, factor)
#check
sapply(df99, class)
skim(df99)
```
#6 Find Nulls
```{r}
#def count_missing:
count_missing = function(df){
  sapply(df, FUN=function(col) sum(is.na(col)))
}
#apply count_missing func to df
nacounts <- count_missing(df99)
hasNA = which(nacounts > 0)
#see results
nacounts[hasNA]
```
  #99 Optional Rename
  ```{r}
  #reassign name
  {new_name} <- df99
  #remove generic name
  remove(df99)
  ```