DATA 607 Project 1 - Chess Tournament Analysis

Understanding the file first

First thing I will do is look at the text file and try to understand what I am working with. This is not a CSV or Excel file. It is just a printed chess tournament table. After watching the chess crosstable video, I understand that: 1. Each player shows up in two lines 2. First line has the player name, total points, and round results 3. Second line has the state and rating information So the file looks messy, but the structure repeats for every player.

Reading the data into R

Next, I will read the text file line by line using R. I will not try to read it as a table, because it is not formatted that way. Reading line by line makes it easier to pull out the parts I need.

Finding where player data starts

Then, I will separate the important lines from the file. I will: 1. Grab the lines that start with a number → these are player lines 2. Grab the lines that start with a state code like MI or ON → these are rating lines This helps me line up each player with their state and rating.

Extracting basic player information

After that, I will extract the basic things I need for each player: • Player name • Player state • Total points • Pre-tournament rating

I will ignore the post-rating because the project only asks for pre-rating.

Getting and Matching opponent information from rounds

Next, I will look at the round results. Each round shows the opponent’s player number, not their name or rating. So what I will do is: Pulling out the opponent numbers from each round and Storing them for each player

Then, I will create a lookup list that connects: Player number → Pre-rating

Using this, I will: 1. Match each opponent number to their pre-rating 2. Collect all opponent ratings for each player

Calculating average opponent rating

After I have all opponent ratings, I will: 1. Take the average of those ratings 2. Ignore games that were not played 3. Round the final average

This gives me the average pre-chess rating of opponents, which is required for the project.

Creating the final dataset

Finally, I will put everything together into one clean dataset with:

• Player Name
• Player State
• Total Points
• Pre-Rating
• Average Opponent Pre-Rating

I will then export this as a CSV file so it can be used in SQL or other systems.

Code Base

library(tidyverse)
library(stringr)

Reading Text File

file_url <- "https://raw.githubusercontent.com/sinemkilicdere/Data607/refs/heads/main/Project1/tournamentinfo.txt"
raw_lines <- readLines(file_url)

Warning in readLines(file_url): incomplete final line found on
'https://raw.githubusercontent.com/sinemkilicdere/Data607/refs/heads/main/Project1/tournamentinfo.txt'

I use readLines() because this is a text file, not a CSV.

Identifying player and rating lines

player_lines <- raw_lines[str_detect(raw_lines, "^\\s*\\d+ \\|")]
state_lines  <- raw_lines[str_detect(raw_lines, "^\\s*[A-Z]{2} \\|")]

Player lines start with a number. State lines start with two capital letters like MI or ON.

Extracting player number, name, and total points

players <- tibble(
  player_num = as.integer(str_extract(player_lines, "^\\s*\\d+")),
  name = str_trim(str_extract(player_lines, "(?<=\\|).+?(?=\\|)")),
  total_points = as.numeric(str_extract(player_lines, "(?<=\\|)\\d+\\.\\d(?=\\s*\\|)"))
)

Extracting player state and pre-rating

states <- str_extract(state_lines, "(?<=^\\s{3})[A-Z]{2}")
ratings <- as.numeric(str_extract(state_lines, "(?<=R:\\s)\\d+"))

players <- players %>%
  mutate(
    state = states,
    pre_rating = ratings
  )

Extracting opponent numbers from each round

get_opponents <- function(line) {
  parts <- str_split(line, "\\|")[[1]]
  round_parts <- parts[4:length(parts)]
  # Extract numbers from each round column
  opp_list <- str_extract_all(round_parts, "\\d+")
  # Flatten list into single numeric vector
  opp_nums <- as.integer(unlist(opp_list))
  opp_nums
}

opponents_list <- lapply(player_lines, get_opponents)

Each round lists the opponent’s player number.

Creating a rating lookup table

rating_lookup <- players %>%
  select(player_num, pre_rating)

Calculate average opponent pre-rating

avg_opponent_rating <- sapply(opponents_list, function(opp_nums) {
  opp_ratings <- rating_lookup$pre_rating[
    match(opp_nums, rating_lookup$player_num)
  ]
  mean(opp_ratings, na.rm = TRUE)
})

players$avg_opponent_rating <- round(avg_opponent_rating)

Final clean dataset

final_df <- players %>%
  select(
    Player_Name = name,
    Player_State = state,
    Total_Points = total_points,
    Pre_Rating = pre_rating,
    Avg_Opponent_Pre_Rating = avg_opponent_rating
  )

final_df

# A tibble: 64 × 5
   Player_Name       Player_State Total_Points Pre_Rating Avg_Opponent_Pre_Rat…¹
   <chr>             <chr>               <dbl>      <dbl>                  <dbl>
 1 GARY HUA          ON                    6         1794                   1605
 2 DAKSHESH DARURI   MI                    6         1553                   1561
 3 ADITYA BAJAJ      MI                    6         1384                   1665
 4 PATRICK H SCHILL… MI                    5.5       1716                   1574
 5 HANSHI ZUO        MI                    5.5       1655                   1588
 6 HANSEN SONG       OH                    5         1686                   1519
 7 GARY DEE SWATHELL MI                    5         1649                   1538
 8 EZEKIEL HOUGHTON  MI                    5         1641                   1468
 9 STEFANO LEE       ON                    5         1411                   1635
10 ANVIT RAO         MI                    5         1365                   1554
# ℹ 54 more rows
# ℹ abbreviated name: ¹​Avg_Opponent_Pre_Rating

write_csv(final_df, "chess_tournament_clean.csv")

Sources

• Chess Crosstable Explanation Video https://youtu.be/T5PXYl2FEUo
• FiveThirtyEight Elo Ratings https://fivethirtyeight.com/tag/elo-ratings/