volleyball-project.Rmd

---
title: "Duke Sports Analtyics: Volleyball Project (Fall 2021)"
description: |
  Exploring Duke Women's Volleyball statistics
author: Hillary Lee, Angel Chaudhary, Aditya Sardesai, Ethan Ouellette
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE, message = FALSE, warning = FALSE,
                      comment = NA, tidy = "styler",
                      fig.width = 8, fig.height = 5)
```

# Introduction

Volleyball is heavily underdeveloped in terms of sports-statistics, which has
prompted us to start this project that aims to develop a WAR-like metric for
volleyball based on statistics from Duke Women's Volleyball for seasons
2018~2021. It will provide a comprehensive and hollistic approach in evaluating
players' strength and performance.

This is an explanation of the data provided on the Duke website based on the
FIVB VIS (Volleyball Information System):

| ATTACK           | Description                                               |
|------------------|-----------------------------------------------------------|
| Kills or Spikes  | Number of scoring attacks.                                |
| Errors or Faults | Number of attacking errors. The opponent scores directly. |
| Attempts         | Total number of attacks.                                  |
| Percent          | Percentage of kills out of attempts.                      |
| Kills Per Set    | Average number of scoring attacks per set.                |

| SET                  | Description                                                  |
|----------------------|--------------------------------------------------------------|
| Assists              | Number of set assists.                                       |
| Assists Per Set      | Number of set assists per set.                               |
| Ball Handling Errors | Number of mistakes in setting. The opponent scores directly. |

| SERVE        | Description                                      |
|--------------|--------------------------------------------------|
| Aces         | Number of points directly scored by the service. |
| Errors       | Number of service mistakes.                      |
| Aces Per Set | Average number of serve aces per set.            |

| RECEPTIONS     | Description                                 |
|----------------|---------------------------------------------|
| Errors         | Number of reception errors.                 |
| Errors Per Set | Average number of reception errors per set. |

| DEFENSE      | Description                     |
|--------------|---------------------------------|
| Digs         | Number of outstanding digs.     |
| Digs Per Set | Average number of digs per set. |

# Scraping Data

```{r packages, message = FALSE}
library(tidyverse)
```

## Duke Data

The data is scraped from Duke Sport's website: https://goduke.com/sports/vb/stats/2021?path=vb.

```{r read-duke-data}
url <- "https://goduke.com/sports/vb/stats/2021?path=vb"
Team_data_2021 <- xml2::read_html(url) %>% 
  rvest::html_table(fill = T) %>% 
  purrr::pluck(1) %>% 
  janitor::clean_names() %>% 
  dplyr::as_tibble() %>% 
  select(-opponents_opp) %>% 
  rename("2021" = duke)

url <- "https://goduke.com/sports/womens-volleyball/stats/2020"
Team_data_2020 <- xml2::read_html(url) %>% 
  rvest::html_table(fill = T) %>% 
  purrr::pluck(1) %>% 
  janitor::clean_names() %>% 
  dplyr::as_tibble() %>% 
  select(-opponents_opp) %>% 
  rename("2020" = duke)

url <- "https://goduke.com/sports/womens-volleyball/stats/2019"
Team_data_2019 <- xml2::read_html(url) %>% 
  rvest::html_table(fill = T) %>% 
  purrr::pluck(1) %>% 
  janitor::clean_names() %>% 
  dplyr::as_tibble() %>% 
  select(-opponents_opp) %>% 
  rename("2019" = duke)

Team_data_compiled <- left_join(left_join(Team_data_2021, Team_data_2020), Team_data_2019)

Team_data_compiled
```

## NCAA Data

The data is scraped from the NCAA website: https://www.ncaa.com/stats/volleyball-women/d1/

```{r read-ncaa-data}
url <- "https://www.ncaa.com/stats/volleyball-women/d1/current/team/48"
NCAA_aces_per_set <- xml2::read_html(url) %>% 
  rvest::html_table(fill = T) %>% 
  purrr::pluck(1) %>% 
  janitor::clean_names() %>% 
  dplyr::as_tibble() %>% 
  rename("aces per set" = per_set,
         "sets" = s)

url <- "https://www.ncaa.com/stats/volleyball-women/d1/current/team/49"
NCAA_blocks_per_set <- xml2::read_html(url) %>% 
  rvest::html_table(fill = T) %>% 
  purrr::pluck(1) %>% 
  janitor::clean_names() %>% 
  dplyr::as_tibble() %>% 
  rename("blocks per set" = per_set,
         "sets" = s)

url <- "https://www.ncaa.com/stats/volleyball-women/d1/current/team/50"
NCAA_digs_per_set <- xml2::read_html(url) %>% 
  rvest::html_table(fill = T) %>% 
  purrr::pluck(1) %>% 
  janitor::clean_names() %>% 
  dplyr::as_tibble() %>% 
  rename("digs per set" = per_set,
         "sets" = s)

url <- "https://www.ncaa.com/stats/volleyball-women/d1/current/team/47"
NCAA_assists_per_set <- xml2::read_html(url) %>% 
  rvest::html_table(fill = T) %>% 
  purrr::pluck(1) %>% 
  janitor::clean_names() %>% 
  dplyr::as_tibble() %>% 
  rename("assists per set" = per_set,
         "sets" = s)

url <- "https://www.ncaa.com/stats/volleyball-women/d1/current/team/46"
NCAA_kills_per_set <- xml2::read_html(url) %>% 
  rvest::html_table(fill = T) %>% 
  purrr::pluck(1) %>% 
  janitor::clean_names() %>% 
  dplyr::as_tibble() %>% 
  rename("kills per set" = per_set,
         "sets" = s)

url <- "https://www.ncaa.com/stats/volleyball-women/d1/current/team/45"
NCAA_hitting_percentage <- xml2::read_html(url) %>% 
  rvest::html_table(fill = T) %>% 
  purrr::pluck(1) %>% 
  janitor::clean_names() %>% 
  dplyr::as_tibble() %>% 
  rename("sets" = s)

NCAA_kills_per_set
NCAA_hitting_percentage
NCAA_aces_per_set

NCAA_blocks_per_set
NCAA_digs_per_set
NCAA_assists_per_set
```

The following is a brief explanation of how the statistics were collected by 
[NCAA regulations](http://fs.ncaa.org/Docs/stats/VB_Forms/Volleyball%20Box%20Score%20Form%20Instructions.pdf).

*Offense*
* Attack
  + A kill is awarded any time an attack attempt is unreturnable by the opposition or any time the attack attempt leads directly to a blocking error by the opponent.
  + Kills per set = total kills / sets
  + Hitting percentage = (kills - errors) / Total attacks
* Service Aces
  + A service ace is a serve that directly results in a point.
  + Aces per set = total team aces / sets

*Defense*
* Blocks
  + Total team blocks are derived by adding all block solos to one half of the block assists.
  + Blocks per set = total team blocks / sets
* Digs
  + A dig is awarded when a player successfully passes a ball that has been attacked by the opposition.
  + Digs are given only when players receive an attacked ball and it is kept in play.
  + Digs per set = total team digs / sets
* Sets
  + An assist is awarded when a player passes, sets or digs a ball to a teammate who attacks the ball for a kill.
  + Assists per set = assists / sets

```{r nitty-gritty}
library(readxl)
nitty_gritties <- read_excel("data/nitty_gritties.xlsx")
nitty_gritties

nitty_gritties_full <- read_excel("data/NCAA_Statistics.xlsx")
nitty_gritties_full
```

This data shows the RPI (rating percentage index) data for the AAC conference, 
where Duke currently ranks 62nd in. The RPI is a quantity used to rank sports 
teams based upon a team's wins and losses and its strength of schedule.

# Data Visualization

## Offense and Defense stats

Section authored by Hillary Lee.

### Offense

Service aces, kills, and hitting percentage are offensive stats that are compiled in this data.

```{r offense}
df1 <- NCAA_kills_per_set %>% select(-c(rank, sets))
df2 <- NCAA_aces_per_set %>% select(-c(rank, sets))
df3 <- NCAA_hitting_percentage %>% select(-c(rank, kills))

compiled_offense <- df1 %>% inner_join(df2, by = "team") %>% inner_join(df3, by = "team")
col_order <- c("team", "sets", "total_attacks", "kills", "errors", "aces", "kills per set", "pct", "aces per set")
compiled_offense <- compiled_offense[, col_order]
compiled_offense
```

### Defense

Blocks, digs, and assists are defensive stats that are compiled in this data.

```{r defense}
df1 <- NCAA_blocks_per_set %>% select(-c(rank, sets))
df2 <- NCAA_digs_per_set %>% select(-c(rank, sets))
df3 <- NCAA_assists_per_set %>% select(-c(rank))

compiled_defense <- df1 %>% inner_join(df2, by = "team") %>% inner_join(df3, by = "team")
col_order <- c("team", "sets", "block_solos", "block_assists", "digs", "assists", "blocks per set", "digs per set",  "assists per set")
compiled_defense <- compiled_defense[, col_order]
compiled_defense
```

## Plotting hitting % vs. win %

Section authored by Ethan Ouellette.

```{r plot-1}

lookup <- nitty_gritties_full[ ,c("Team", "Adj. RPI Value")]
hitting_pct <- NCAA_hitting_percentage[, c("team", "pct")]
hitting_pct_rpi <- merge(lookup, hitting_pct, by.x="Team", by.y="team")
names(hitting_pct_rpi)[2] <- "rpi"

hitting_pct_rpi

ggplot(data = hitting_pct_rpi, mapping = aes(x = pct, y = rpi)) + 
  geom_point() +
  labs(title = paste("Hitting Percentage vs. Rating Percentage Index"))
```

```{r linear model}
ncaa_and_nitty <- left_join(nitty_gritties_full, NCAA_assists_per_set, 
                             by = c("Team" = "team")) %>%
  left_join(NCAA_aces_per_set, by = c("Team" = "team")) %>%
  left_join(NCAA_blocks_per_set, by = c("Team" = "team")) %>%
  left_join(NCAA_digs_per_set, by = c("Team" = "team")) %>%
  left_join(NCAA_hitting_percentage, by = c("Team" = "team")) %>%
  left_join(NCAA_kills_per_set, by = c("Team" = "team"))

ncaa_and_nitty
```

rpi_lm <- linear_reg() %>%
  set_engine("lm") %>%
  fit(Adj. RPI Value ~ per_set, data = ncaa_and_nitty)
  
## Correlation between Home/Away Games and Winning Percentage

Section authored by Aditya Sardesai.

# Analysis

This section will be an analysis.

# Summary

This section will be a summary of the report.