This and any future tutorials are not intended to be construed as:

An endorsement of R as the best programming language.

A characterization of Python as a bad or inferior language.

As I detailed in my Sports Illustrated Hackathon keynote presentation, it doesn’t matter which programming language you pick so long as you master it to and use your skills to solve problems.

Any of the major programming languages {R, Python, JavaScript, C, ect..} with some expertise empower you to do the truly amazing things that programming languages can do! If you are a true amateur I’d advocate taking some time to learn about the major languages and then deciding which you think is best suited for you to learn.

That said, this and future tutorials are intended to destroy some misconceptions about the language that I love, R. There are myths you may come across that imply R is inferior to Python for web-scraping, that it’s syntax doesn’t make sense and that the language is too hard to learn.

None of these things are true and I hope by the end of this post, whatever your existing programming language of choice is, learned or intended to be learned, you won’t disagree with this statement.

Enough of the Boring Stuff, Let’s Gokul

In order to complete today’s tutorial you are going to ensure your computer is locked and load with a few things.

First make sure you have R and RStudio installed.

Then if you don’t already use Firefox or Chrome pick a browser install it.

Finally, launch your chosen browse and install the fantastic SelectorGadget widget.

Fire up

It’s game time.

The final step before we are ready for action is to ensure you have the necessary R packages installed and loaded. I am using the development version of many of these packages, so a word caution there.

For those who need to install the packages, I’ve referenced the github repos in the comments, just copy and paste the code starting at devtools:: until the end the line into the console or in a script that you can run. For the non development packages just copy and paste starting from install.packages and do the same as before.

packages <- #These are the packages we need
  c(
    'devtools', # install.packages(devtools)
    'dplyr', # devtools::install_github('hadley/dplyr')
    'magrittr', # devtools::install_github('smbache/magrittr')
    'rvest', # devtools::install_github('hadley/rvest')
    'data.table', # install.packages(data.table)
    'lubridate', # devtools::install_github(hadley/lubridate)
    'DataCombine', #devtools::install_github('christophergandrud/DataCombine')
    'stringr', # devtools::install_github('hadley/stringr')
    'readr', # devtools::install_github('hadley/readr')
    'formattable', #devtools::install_github('renkun-ken/formattable')
    'tidyr' # devtools::install_github('hadley/tidyr')
  )

lapply(packages, library, character.only = T)

In order to mine for NBA draft data gold we need to know a few things, the most important of which being what years has the draft has existed?

To answer this question we navigate our browsers here http://www.basketball-reference.com/draft/. It looks like the draft has existed since 1947 if you include the NBA’s precursor league the BAA and 1950 if you don’t.

We are almost ready to start mining we just need to pick a draft year to test. Given that it was the year this distinguished author’s birth, let’s go with 1983. From the main draft index page click on the 1983 Draft link.

Look at all that beautiful data just waiting for us to suck in, I am giddy.

However, before we get ahead of ourselves thinking we’ve unlocked the magic formula to bringing data we long to explore into R we must find one more key input. This important missing piece is the CSS identifier of the data we want from the page, in this case the draft table. Think of this piece of information as unique identifier for the the html data on the page we want.

Remember that SelectorGadget I mentioned earlier? It’s time to use it to extract the CSS identifier we need.

Click the widget’s button.

Scroll over the table with the draft information

Click once an orange box surrounds the table

You should see something like the picture below

Just like the 1978 Cars song said, we’ve got Just What We Needed it’s time to go data mining.

draft_year <- # our chosen draft year
  1983

year.first_draft <- # remember from earlier 1st BAA draft
  1947

current_month <- # we're gonna to teach the code to know if the draft has
  Sys.Date() %>%
  month %>%
  as.numeric

if (current_month > 6) {
  year.most_recent_draft <- # the draft is at the end of June so if its July we've passed the draft!
    Sys.Date() %>% # tells us what today is
    year %>% # extracts the year
    as.numeric()
} else {
  year.most_recent_draft <-
    Sys.Date() %>%
    year %>%
    as.numeric() - 1 #if it's earlier than July we take last year's draft
}

if (!draft_year %in% year.first_draft:year.most_recent_draft) {
  stop.message <-
    "Not a valid draft year boss!!  Drafts can only be between " %>%
    paste0(year.first_draft, ' and ', year.most_recent_draft)
  stop(stop.message)
}

if (draft_year < 1950) {
  base <- # remember what we learned earlier?
    'http://www.basketball-reference.com/draft/BAA_'

  id.league <-
    'BAA' # based on the date R will know hte league
} else {
  base <- # well if it's not the BAA it's the NBA, DUHHH
    'http://www.basketball-reference.com/draft/NBA_'

  id.league <-
    'NBA'
}

url.draft_year <-
  base %>%
  paste0(draft_year, '.html') # creates the url where the data lives

url.draft_year # our url with the data and it should work for any draft year!

## [1] "http://www.basketball-reference.com/draft/NBA_1983.html"

page <- # Get the html from the page into R
  url.draft_year %>%
  read_html

raw_data <-
  page %>%
  html_nodes('#stats') %>% # Remember this is the css id from earlier
  html_table(header = F, fill = F) %>% # This function reads the table
  data.frame %>% # Data is not ts not a data.frame this puts it into tht form
  tbl_df # This converts it into a super dope special type of data fram tbl_df

headers <- # Get the parent header rows so we can append them into the column names and sove the duplicate name problem
  raw_data %>%
  slice(1) %>% # Takes the 1st row where this header info
  unlist %>% # Returns a the row as list we dont want that
  as.character %>% # We want to explicitly define this vector as a character
  tolower %>% # Gonna be part of our title, I perfer lower case titles
  str_replace('\\ ', '_') # Column names SHOULD NOT contain spaces, use the snake!!

# Time to get the actual column items

columns <-
  raw_data %>%
  slice(2) %>% # This information lives in the second row
  tolower %>%
  str_replace('%', '_pct') %>% # % should also never be a column header this gets rid of them
  str_replace('/', '_per_') # / should ALSO never be a column header, this removes them

name.df <- # This creates a data frame that will contain column names
  data_frame(header = headers, column = columns) %>% # data_frame is a special faster data frame from dplyr that I reccomend you use whever possible
  mutate( # Mutate adds variables to data frames
    header = ifelse(header == '', NA, header), # No headers for blank items, they happen to be identifiers about the draft player versus statistics
    header = ifelse(header %like% 'round|territorial_picks', NA, header) # If we see these words we want to exclude that row
  ) %>%
  FillDown('header') %>% # Fills down the keys
  mutate(name.column = ifelse(header %>% is.na, column, paste(header, column, sep = '.')) # if the field has a header we want to join the items together if not just keep the item as the name, we know which those are by where the NAs live
         )

names(raw_data) <- # change the names to our new clean names!
  name.df$name.column

## Use R skills to magically figure out which round the player was taken in

round_rows <- # find out the row number where we get indication of the round
  'Round' %>% # that is the word we need to look for
  grep(raw_data$player) # we want to return the row numbers in the 4th column where we find that word, the round headers live in this column

round_df <- #create data frame where the rounds start
  data_frame(round = paste0('Round ', 1:length(round_rows)), # the number of times we see the word indicates how many rounds that year had with some slight exceptions
             id.row = round_rows + 2 # the players names start 2 rows below where we find the word
             )

## This is for later but trust me from exploring, there were other ways players ended up on teams that weren't considered drafts, there were specific words that Basketball-Reference uses to delinate them and those are the 2 words we looking for

if ('Other|Territorial Picks' %>% grep(raw_data$player) %>% #looking for their existence, if they exist the length of the results won't be zero
    length > 0) {
  other_rows <-
    'Other|Territorial Picks' %>% grep(raw_data$player)

  other_rounds.df <-
    data_frame(round = 'Other', #create a special name so we will get an NA for the round when we magically extract the round later
               id.row = other_rows + 2)
  round_df %<>%
    bind_rows(other_rounds.df)
}

raw_data %<>%
  mutate(id.row = 1:nrow(.)) %>% # We need to create a temporary id to merge the rows in the draft data frame to this, here we do this
  left_join(round_df, by = 'id.row') %>% # A left join keeps the first table and merges any matching table from the right table, in this case the right table is the round data, frame
  select(round, everything()) %>% # cleans up order
  FillDown('round') %>% # Fills the round by the boundries
  mutate(id.round = round %>% extract_numeric) %>% # We want to extract the numeric round if it exists this does that
  select(id.round, everything()) %>% # cleans up the order again
  select(-c(id.row, round)) # we dont need these columns, using - tells R to remove these columns

## Last step, remove the rows we don't need

raw_data %<>% ## remove the rows we don't need
  slice(-1) %>% # First row the way we import the data contains the items, remember
  dplyr::filter(!rk == 'Rk', !player %like% 'Round|Other Picks') # In these tables we know how to find "bad" rows by looking for "Rk" in the rk column and the word Round or Other Picks in the player column filters out fields that don't contain data because they contain those 3 words

raw_data %<>%
  select(-rk) %>% # We dont need the rank column its meaningless
  rename( #A little more name cleaning
    id.pick = pk, # better name, helps us remember that the pick is an id which it is
    id.bref.team = tm, # this column corresponds with the basketball reference
    totals.years_played = yrs # Better name, shows us how many years someone played
  ) %>%
  arrange(id.pick) #orders by pick!

numeric_columns <- # Tell R which fields are numeric and then convert them to numeric fields!
  raw_data %>%
  select(-c(id.bref.team, college, player)) %>% # Selects the non numeric fields
  names # Returns the names of the columns that we want to convert

raw_data %<>%
  mutate_each_(funs(as.numeric),  # this is where we tell R what function we want to use on our selected columns, here its as.numeric
               vars = numeric_columns # These are the columns from earlier
  )

raw_data %>%
  glimpse # This function gives us a snapshot of our data frame!

## Observations: 226
## Variables: 22
## $ id.round            (dbl) 1, 1, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 1, 5, ...
## $ id.pick             (dbl) 1, 10, 100, 101, 102, 103, 104, 105, 106, ...
## $ id.bref.team        (chr) "HOU", "WSB", "UTA", "DET", "DAL", "WSB", ...
## $ player              (chr) "Ralph Sampson", "Jeff Malone", "Matt Clar...
## $ college             (chr) "University of Virginia", "Mississippi Sta...
## $ totals.years_played (dbl) 9, 13, NA, 1, 1, NA, NA, NA, NA, NA, NA, N...
## $ totals.g            (dbl) 456, 905, NA, 7, 1, NA, NA, NA, NA, NA, NA...
## $ totals.mp           (dbl) 13591, 29660, NA, 28, 16, NA, NA, NA, NA, ...
## $ totals.pts          (dbl) 7039, 17231, NA, 12, 3, NA, NA, NA, NA, NA...
## $ totals.trb          (dbl) 4011, 2364, NA, 3, 5, NA, NA, NA, NA, NA, ...
## $ totals.ast          (dbl) 1038, 2154, NA, 1, 0, NA, NA, NA, NA, NA, ...
## $ shooting.fg_pct     (dbl) 0.486, 0.484, NA, 0.462, 0.333, NA, NA, NA...
## $ shooting.3p_pct     (dbl) 0.172, 0.268, NA, NA, NA, NA, NA, NA, NA, ...
## $ shooting.ft_pct     (dbl) 0.661, 0.871, NA, NA, 0.500, NA, NA, NA, N...
## $ per_game.mp         (dbl) 29.8, 32.8, NA, 4.0, 16.0, NA, NA, NA, NA,...
## $ per_game.pts        (dbl) 15.4, 19.0, NA, 1.7, 3.0, NA, NA, NA, NA, ...
## $ per_game.trb        (dbl) 8.8, 2.6, NA, 0.4, 5.0, NA, NA, NA, NA, NA...
## $ per_game.ast        (dbl) 2.3, 2.4, NA, 0.1, 0.0, NA, NA, NA, NA, NA...
## $ advanced.ws         (dbl) 20.1, 54.2, NA, 0.0, 0.0, NA, NA, NA, NA, ...
## $ advanced.ws_per_48  (dbl) 0.071, 0.088, NA, -0.046, -0.003, NA, NA, ...
## $ advanced.bpm        (dbl) 0.0, -2.1, NA, -6.9, -12.1, NA, NA, NA, NA...
## $ advanced.vorp       (dbl) 6.8, -0.9, NA, 0.0, 0.0, NA, NA, NA, NA, N...

player <-
  page %>%
  html_nodes('td:nth-child(4) a') %>% # This is the column where the player IDs live
  html_text %>% # this takes the html output and returns text
  str_trim # To be safe this trims the code in case there are any unceassary white spaces

url.bref.player <- # This creates the player URL
  page %>%
  html_nodes('td:nth-child(4) a') %>%
  html_attr('href') %>% # This function pulls in the html attribute, in this case the stem
  paste0('http://www.basketball-reference.com', .) # We need to append the base URL!

stem.player <- # Here we are going to extract out from the steam the exact player ID
  page %>%
  html_nodes('td:nth-child(4) a') %>%
  html_attr('href') %>%
  str_replace_all('/players/|.html', '') # eliminates the unnecassary words to help us get at the clean ID

players_urls <- # This will create a data frame with the information we want
  data_frame(player, url.bref.player, stem.player) %>%
  separate(stem.player, c('letter.first', 'id.bref.player'), #  Separates the remaining 2 parts by its delimiter the / we only want the second column which contains the id
           sep = '/') %>%
  select(-letter.first) # removes the unneeded column

## Resolve the team ids

teams_ids <-
    'https://asbcllc.com/data/nba/bref/nba_teams_ids.csv' %>%
    read_csv # imports my team data

data <- #create a new data frame which will be our final data frame with all the information
  raw_data %>%
  left_join(players_urls, by = 'player') %>% # joins are existing data with the players who have BREF profiles
  mutate(id.league, # add in the league id from earlier
         year.draft = draft_year, # add in the draft year from earlier, I like to use periods to separate columns, total personal preference
         url.bref.draft = url.draft_year # this tells us the URL where we got our data from
         )

data %<>%
  left_join(teams_ids, by = 'id.bref.team') %>% # Joins our data frame with the team data which will merge in the missing team data
  select( #reorder the data
    id.league,
    year.draft,
    id.round,
    id.pick:player,
    id.bref.player,
    id.bref.team,
    team,
    id.bref.current_team,
    current_team,
    everything() #select everything else in the data frame
  ) %>%
  arrange(id.pick)

### Time to Messagify!!

players <-
  data %>%
  nrow #counts the number of players in that draft

random_player <- # I want our message to return information about a random player from that draft
  data %>%
  arrange(desc(totals.pts)) %>% # arranges the data with the top scorers first
  mutate(rank.total_points = 1:nrow(.)) %>% # adds a ranking field
  dplyr::filter(totals.pts > 0 & totals.years_played > 1) %>% #filters for players who played over 1 season and scored points
  dplyr::filter(!is.na(id.bref.player)) %>% # filters out players who don't have a BREF profile page
  sample_n(1) # returns 1 random row with a player that fits the criteria!

# Print the message
"Congratulations you pulled in data for all " %>%
  paste0(
    players,
    ' players from the ',
    draft_year,
    ' Draft\nHave you heard of the #',
    random_player$id.pick,
    ' pick, ',
    random_player$player,
    '?\nHe played ',
    random_player$totals.years_played,
    ' seasons & ranked #',
    random_player$rank.total_points,
    ' in his draft class for total points scored!'
  ) %>%
  message

## Congratulations you pulled in data for all 226 players from the 1983 Draft
## Have you heard of the #33 pick, Dirk Minniefield?
## He played 3 seasons & ranked #33 in his draft class for total points scored!

data %>%
  select(id.round, id.pick, player, team, totals.years_played, totals.pts) %>%
  mutate(totals.pts = totals.pts %>% comma(digits = 0)) %>% # formats to add a comma
  head %>% # Takes the top 5 players
  format_table(align = 'c') # This creates the beautiful SVG table you are looking at

get_nba_year_draft_data <-
  function(draft_year = 1983, # this is the default year if you run the function with no year sepcified
           return_message = T # this specifies the default behavior is to return our message
           ) {
    options(warn = -1) # Turns off some annoying warnings

    #Make Smart Draft Years

    year.first_draft <-
      1947

    current_month <-
      Sys.Date() %>%
      month %>%
      as.numeric

    if (current_month > 6) {
      year.most_recent_draft <-
        Sys.Date() %>%
        year %>%
        as.numeric()
    } else {
      year.most_recent_draft <-
        Sys.Date() %>%
        year %>%
        as.numeric() - 1
    }

    if (!draft_year %in% year.first_draft:year.most_recent_draft) {
      stop.message <-
        "Not a valid draft year boss!!  Drafts can only be between " %>%
        paste0(year.first_draft, ' and ', year.most_recent_draft)
      stop(stop.message)
    }

    if (draft_year < 1950) {
      base <-
        'http://www.basketball-reference.com/draft/BAA_'

      id.league <-
        'BAA'
    } else {
      base <-
        'http://www.basketball-reference.com/draft/NBA_'

      id.league <-
        'NBA'
    }

    url.draft_year <-
      base %>%
      paste0(draft_year, '.html')

    page <-
      url.draft_year %>%
      read_html

    raw_data <-
      page %>%
      html_nodes('#stats') %>%
      html_table(header = F, fill = F) %>%
      data.frame %>%
      tbl_df

    headers <-
      raw_data %>%
      slice(1) %>%
      unlist %>%
      as.character %>%
      tolower %>%
      str_replace('\\ ', '_')

    #Time to get the column items

    columns <-
      raw_data %>%
      slice(2) %>%
      tolower %>%
      str_replace('%', '_pct') %>%
      str_replace('/', '_per_')

    name.df <-
      data_frame(header = headers, column = columns) %>%
      mutate(
        header = ifelse(header == '', NA, header),
        header = ifelse(header %like% 'round|territorial_picks', NA, header)
      ) %>%
      FillDown('header') %>%
      mutate(name.column = ifelse(header %>% is.na, column,
                                  paste(header, column, sep = '.'))
             )

    names(raw_data) <-
      name.df$name.column

    ## Magically figure out which round the player was taken in

    round_rows <-
      'Round' %>%
      grep(raw_data$player)

    round_df <-
      data_frame(round = paste0('Round ', 1:length(round_rows)),
                 id.row = round_rows + 2
      )

    if ('Other|Territorial Picks' %>% grep(raw_data$player) %>% length > 0) {
      other_rows <-
        'Other|Territorial Picks' %>% grep(raw_data$player)

      other_rounds.df <-
        data_frame(round = 'Other',
                   id.row = other_rows + 2)
      round_df %<>%
        bind_rows(other_rounds.df)
    }

    raw_data %<>%
      mutate(id.row = 1:nrow(.)) %>%
      left_join(round_df, by = 'id.row') %>%
      select(round, everything()) %>%
      FillDown('round') %>%
      mutate(id.round = round %>% extract_numeric) %>%
      select(id.round, everything()) %>%
      select(-c(id.row, round))

    raw_data %<>%
      slice(-1) %>%
      dplyr::filter(!rk == 'Rk', !player %like% 'Round|Other Picks')

    raw_data %<>%
      select(-rk) %>%
      rename(
        id.pick = pk,
        id.bref.team = tm,
        totals.years_played = yrs
      ) %>%
      mutate(id.pick = id.pick %>% as.numeric) %>%
      arrange(id.pick)

    numeric_columns <-
      raw_data %>%
      select(-c(id.bref.team, college, player)) %>%
      names

    raw_data %<>%
      mutate_each_(funs(as.numeric),
                   vars = numeric_columns
      )

    ## Player ID Extraction
    player <-
      page %>%
      html_nodes('td:nth-child(4) a') %>%
      html_text

    url.bref.player <-
      page %>%
      html_nodes('td:nth-child(4) a') %>%
      html_attr('href') %>%
      paste0('http://www.basketball-reference.com', .)

    stem.player <-
      page %>%
      html_nodes('td:nth-child(4) a') %>%
      html_attr('href') %>%
      str_replace_all('/players/|.html', '')

    players_urls <-
      data_frame(player, url.bref.player, stem.player) %>%
      separate(stem.player, c('letter.first', 'id.bref.player'),
               sep = '/') %>%
      select(-letter.first)

    ## resolve with team
    teams_ids <-
      'https://asbcllc.com/data/nba/bref/nba_teams_ids.csv' %>%
      read_csv

    data <-
      raw_data %>%
      left_join(players_urls, by = 'player') %>%
      mutate(id.league,
             year.draft = draft_year,
             url.bref.draft = url.draft_year)

    data %<>%
      left_join(teams_ids, by = 'id.bref.team') %>%
      select(
        id.league,
        year.draft,
        id.round,
        id.pick:player,
        id.bref.player,
        id.bref.team,
        team,
        id.bref.current_team,
        current_team,
        everything()
      ) %>%
      arrange(id.pick)

    if (
      return_message == T # this will only run our message if we tell it to in our function
        ) {
      players <-
        data %>% nrow

      random_player <-
        data %>%
        dplyr::filter(totals.pts > 0 & totals.years_played > 1) %>%
        arrange(desc(totals.pts)) %>%
        mutate(rank.total_points = 1:nrow(.)) %>%
        dplyr::filter(!is.na(id.bref.player)) %>%
        sample_n(1)

      "Congratulations you pulled in data for all " %>%
        paste0(
          players,
          ' players from the ',
          draft_year,
          ' Draft\nHave you heard of the #',
          random_player$id.pick,
          ' pick, ',
          random_player$player,
          '?\nHe played ',
          random_player$totals.years_played,
          ' seasons & ranked #',
          random_player$rank.total_points,
          ' in his draft class for total points scored!'
        ) %>%
        message
    }
    return(data) # returns the final data frame, this is what we see!
  }

data <-
  get_nba_year_draft_data(draft_year = 2013,return_message = T)

## Congratulations you pulled in data for all 60 players from the 2013 Draft
## Have you heard of the #34 pick, Isaiah Canaan?
## He played 2 seasons & ranked #22 in his draft class for total points scored!

data %>%
  select(id.round, id.pick, player, team, totals.years_played, totals.pts, per_game.pts, per_game.trb, per_game.ast, advanced.ws) %>%
  mutate(totals.pts = totals.pts %>% comma(digits = 0)) %>% #some formating
  head(15) %>%  # Takes the top 15
  format_table(align = 'c')

get_all_nba_draft_data <- function(year_start = NA, year_end = NA,
                                   include_baa = T, return_message = T) {

  # This makes sure the first function is in your environment
  if ('get_nba_year_draft_data' %in% ls() ){
    #
  }

  if (
    include_baa == T # asks us if we want BAA data, assumes we do
    ) {
    year.first_draft <-
      1947 # first BAA draft year
  } else {
    year.first_draft <-
      1950 # first NBA draft year
  }

  if (year_start %>% is.na) {
    year_start <-
      year.first_draft # takes the first BAA or NBA draft year if we don't give our code a start year
  }

  current_month <-
    Sys.Date() %>%
    month %>%
    as.numeric

  if (current_month > 6) {
    year.most_recent_draft <-
      Sys.Date() %>%
      year %>%
      as.numeric()
  } else{
    year.most_recent_draft <-
      Sys.Date() %>%
      year %>%
      as.numeric() - 1 # Takes the most recent draft if we don't identify an explicit end to the draft range
  }

  if (year_end %>% is.na) {
    year_end <-
      year.most_recent_draft #If we don't select an end draft year overwrite with the most recent
  }

  draft_years <-
    year_start:year_end #create a numeric vector of the draft years we want

  ## Loop through the selected draft years, get the data and append it to the master data frame

  all_data <-
    data_frame() # form an empty data_frame that will act as the master data frame

  for (year in draft_years){
    data <-
      get_nba_year_draft_data(draft_year = year, return_message = F) # we dont want to see a bunch of messages so lets turn it off for this

    all_data %<>% # takes the master data frame
      bind_rows(data) # binds the rows of the year's data frame to that
  }

  if (return_message == T) {
    players <-
      all_data %>%
      nrow

    ## We want this message to show us some random facts about one of the top 1000 scorers

    random_player <-
      all_data %>%
      dplyr::filter(totals.pts > 0) %>% #people who have scored
      arrange(desc(totals.pts)) %>% #sort by points
      mutate(rank.total_points = 1:nrow(.)) %>%
      dplyr::filter(!is.na(id.bref.player)) %>%
      slice(1:1000) %>%  #take the top 1000s players by points
      sample_n(1) #take a sample of 1

    "Congratulations you pulled in data for " %>%
      paste0(
        players,
        ' players from the ',
        year_start,
        ' to ',
        year_end,
        ' drafts\nHave you heard of the #',
        random_player$id.pick,
        ' pick in the ',
        random_player$year.draft,
        ' Draft, ',
        random_player$player,
        '?\nHe played ',
        random_player$totals.years_played,
        ' seasons & ranks #',
        random_player$rank.total_points,
        ' all time in total points scored during your selected draft eras!'
      ) %>%
      message
  }
  return(all_data) #returns the data
}

all_data <-
  get_all_nba_draft_data()

## Congratulations you pulled in data for 7789 players from the 1947 to 2015 drafts
## Have you heard of the #4 pick in the 1985 Draft, Xavier McDaniel?
## He played 12 seasons & ranks #170 all time in total points scored during your selected draft eras!

all_data %>%
  select(year.draft, id.round, id.pick, player, team, totals.years_played, totals.pts, per_game.pts, per_game.trb, per_game.ast, advanced.ws) %>%
  arrange(desc(totals.pts)) %>%
  mutate(totals.pts = totals.pts %>% comma(digits = 0)) %>% #some formating
  head(15) %>% #lets do the top 15 picks
  format_table(align = 'c')

all_data %>%
  write_csv(path = 'data/nba_baa_drafts_1947_2015.csv') #pick where ever you want to save it this is just my choice.

We did it ladies and gentleman. We now have a function that can pull in any and all NBA or BAA drafts we give it.

Today’s tutorial ends here but we now have refined NBA draft data gold ready for all sorts of interesting data analysis and visualization, things R also excels at.

Check back soon and for a tutorial that will teach how how to visualize and analyze this dataset.

I hope you enjoyed this and recognize that R can do anything that Python does with ease and grace.

Also, if you ever hear someone say R isn’t good at web-scraping you can shake your head and laugh!

You can find the code for the functions here and I urge you to download or fork the repo and use the code whenever you need it!

As always please don’t hesitate to reach out to me on twitter with any questions, comments and concerns!