FunctionList.Rmd

```{r}
#zip codes for state
data(zipcode)

#takes a value for state and returns all the zipcodes from the state
zips_from_state<-function(state_name){
  zip_holder<-zipcode%>%
    filter(state==state_name)%>%
    select(zip)
  zip_state<-zip_holder[,1]
  return(zip_state)
}


```
```{r}
#groups by diferent units
group_by_unit<-function(data, unit){
  if(unit=="zipcode"){
    countbyzip<-data%>%
    group_by(zipcode) %>% 
    count() %>%
    arrange(n) 
    
    return(countbyzip)
  } else if (unit==="county") {
    countbycounty<-data%>%
    group_by(county) %>% 
    count() %>%
    arrange(n) 
    
    return(countbycounty)
  } else if (unit=="state"){
    countbystate<-data%>%
    group_by(state) %>% 
    count() %>%
    arrange(n) 
    
    return(countbystate)
  } else if (unit=="city"){
    countbycity<-data%>%
    group_by(city) %>% 
    count() %>%
    arrange(n) 
    
    return(countbycity)
  } else if (unit=="type"){
    countbyprovider<-data%>%
    group_by(Primary_Taxonomy) %>% 
    count() %>%
    arrange(desc(n)) 
    
    return(countbyprovider)
  }
}

```
```{r}
countbyzip <- provider.data %>% #creating a count of practices by zip code
group_by(zipcode) %>% 
count() %>%
arrange(n) 

#countbyzip
#create a bar graph of providers by frequency
plot<-ggplot(countbyzip, aes(x=zipcode, y=n, fill="#FF6666"))+
  geom_bar(stat="identity")
plot

#number of counties with 1 provider, 2 providers, etc

number_practices<-countbyzip %>%
  select(n) %>% #selecting and grouping by frequency
  group_by(n) %>%
  count() %>%
  arrange(n) #arranging by frequency
number_practices.low <- head(number_practices) #getting the first rows of the number_pratices and desingnating them as low numbers of practices

#add axis names
plot<-ggplot(number_practices.low, aes(x=factor(n), y=nn))+
  geom_bar(stat="identity") + labs(x = "Number of providers", y = "Number of zip codes")
plot
```

```{r}
#Graphs of Provider Type
provider_grouping<-provider.data %>% #creating a count of practices by providers
group_by(Primary_Taxonomy) %>% 
count() %>%
arrange(desc(n)) 
#provider_grouping

#general scatter of all providers types
provider_scatter<-ggplot(provider_grouping, aes(x=Primary_Taxonomy, y=n))+
 geom_point()+
    theme(axis.text.x=element_blank(),
        axis.ticks.x=element_blank())+
  labs(x = "Provider Types", y = "Number of providers")+
  scale_color_brewer(palette="Set1")
provider_scatter

#create a bar graph of top 5 providers by frequency
top5providers<-head(provider_grouping)
providers_bar<-ggplot(top5providers, aes(x=Primary_Taxonomy, y=n))+
  geom_bar(stat="identity")+
  theme(axis.text.x = element_text(angle=90))+
  labs(x = "Provider Type", y = "Number of Providers")
providers_bar

```

```{r}
NumProviderInZip<-function(zipcode,taxonomy){
  url1<- "https://npiregistry.cms.hhs.gov/registry/search-results-table?addressType=ANY&postal_code=" #setting the url to scrape from
  provider.data <- data.frame() #initializing an empty data frame
  skips <- seq(0,9999999,100) #create skips
  for (i in 1:length(zipcode)) { #iterating over all RI zip codes
    for (j in 1:length(skips)){ #also iterating over skils
      zip <- zipcode[i]
      skip <- skips[j]
      tax <- str_replace_all(taxonomy," ","+")
      url<-paste0(url1,zip,"&skip=",skip,"&taxonomy_description=",tax) #pasting the url, with the rhode island zip code and including the skips
      #text scrape to pull our places by zip code
      h <- read_html(url, timeout = 200)
      reps <- h %>% #setting up the repeating structure
        html_node("table") %>%
        html_table()
      if (any(is.na(reps[,1])) | all(is.na(reps[,1]))) { #if the page has no physicians, move to next zip code
        break}
      reps$zip <- zipcode[i]
      provider.data <- rbind(provider.data,reps) #binding together the provider data and reps data
    }
  }
  colnames(provider.data) <- c("NPI","Name","NPI_Type","Primary_Practice_Address","Phone","Primary_Taxonomy","zipcode")

  provider.data$street <- noquote( str_extract(provider.data$Primary_Practice_Address,pattern="^[a-zA-Z0-9_. -]+?(?=\n)")) #street name and number
  provider.data$city <- noquote( str_extract(provider.data$Primary_Practice_Address,pattern="(?<=\t)[a-zA-Z_. -]+?(?=, )"))#city
  provider.data$statename<- noquote( str_extract(provider.data$Primary_Practice_Address,pattern="(?<=, )[A-Z]+(?=\\s)")) #state postal code (2 characters)
  provider.data<-mutate(provider.data, zipcode= as.character(zipcode))

  zip_link<-read.csv("zcta_county_rel_10.txt") %>%
    select(ZCTA5, STATE, COUNTY, GEOID) %>%
    rename(zipcode = ZCTA5) %>%
    mutate(zipcode = as.character(zipcode))
  zip_link$zipcode = stri_pad_left(zip_link$zipcode, 5, "0")

  NPI_join<-inner_join(provider.data, zip_link, by="zipcode")
  census<-read.csv("co-est2017-alldata.csv")

  NPI_to_census<-inner_join(NPI_join, census, by=c("STATE", "COUNTY"))
  return(nrow(NPI_to_census))
}
```
```{r}




```