s2-exercises-with-answers.Rmd

---
title: "Session 2 Exercises (with answers)"
author: "Ben Raymond, Adrien Ickowicz"
output:
  html_document:
    toc: false
    theme: flatly
    highlight: zenburn
    css: "extra/exercises.css"
---

```{r chunkopts, eval = TRUE, echo = FALSE}
knitr::opts_chunk$set(message = FALSE, warning = FALSE)
options(width=80)
options(tibble.width = 80, tibble.print_max = 10, tibble.print_min = 10)

```

# Preparation

Load some packages:

```{r setup}
library(tidyverse)
library(datavolley)
library(ovlytics)

```

Read in our data

```{r data}
f <- dir("example_data/DE Men 2019", full.names = TRUE, pattern = "\\.dvw$")
x <- lapply(f, dv_read, skill_evaluation_decode = "german")

## concatentate the play-by-play data into one big data frame
px <- bind_rows(lapply(x, plays))

```

Check for inconsistent team names / IDs:

```{r check1}
px %>% count(team, team_id) %>% arrange(team)

x <- remap_team_names(x, remap = tibble(from = "BERLIN", to = "BERLIN RECYCLING Volleys", to_team_id = "5413"))

px <- bind_rows(lapply(x, plays))
px %>% count(team, team_id) %>% arrange(team)

check_player_names(x)

```

Add some extra columns:

```{r extracols}
px <- ov_augment_plays(px, to_add = "all")

```


# Exercises

### 1. What is the mean serve error rate?

<details>
  <summary>Click for solution</summary>

```{r}
px %>% filter(skill == "Serve") %>%
    summarize(error_rate = mean(evaluation == "Error"))

```

</details>


### 2. By team?

<details>
  <summary>Click for solution</summary>

```{r}
px %>% filter(skill == "Serve") %>%
    group_by(team) %>%
    summarize(error_rate = mean(evaluation == "Error"))

```

</details>

### 3. By team, but only for Herrsching and Giesen? (First find their full team names in the data set)

<details>
  <summary>Click for solution</summary>

```{r}
px %>% filter(skill == "Serve") %>%
    filter(team %in% c("HELIOS GRIZZLYS Giesen", "WWK Volleys Herrsching")) %>%
    group_by(team) %>%
    summarize(error_rate = mean(evaluation == "Error"))

```

</details>

### 4. By player?

<details>
  <summary>Click for solution</summary>

```{r}
px %>% filter(skill == "Serve") %>%
    group_by(player_name) %>%
    summarize(error_rate = mean(evaluation == "Error"))

```

Followup: would it be helpful to `group_by(player_name, player_id)` or `group_by(player_name, team)` ?

</details>


### 5. Produce a table (data frame) of mean serve error rate and mean ace rate by team

<details>
  <summary>Click for solution</summary>

```{r}
team_errors_aces <- px %>% filter(skill == "Serve") %>%
    group_by(team) %>%
    summarize(error_rate = mean(evaluation == "Error"), ace_rate = mean(evaluation == "Ace"))

team_errors_aces

```

### 6. Use that to plot mean serve error rate vs mean ace rate

<details>
  <summary>Click for solution</summary>

```{r}
ggplot(team_errors_aces, aes(error_rate, ace_rate)) + geom_point()

```

</details>


### 7. What is the average sideout rate by pass quality (i.e. sideout rate on perfect passes, sideout rate on good passes, etc)

<details>
  <summary>Click for solution</summary>

```{r}
px %>% filter(skill == "Reception") %>%
    group_by(evaluation) %>%
    summarize(sideout_rate = mean(point_won_by == team))

```

</details>


### 8. Calculate the first-ball attack kill rate (i.e. kill rate on attacks made on serve reception)

<details>
  <summary>Click for solution</summary>

```{r}
px %>% filter(skill == "Attack" & phase == "Reception") %>%
    group_by(team) %>%
    summarize(first_ball_kill_rate = mean(evaluation == "Winning attack"))

```

</details>


### 9. Calculate the breakpoint rate (fraction of points won while serving) by rotation (use the `setter_position` column), for all teams combined

<details>
  <summary>Click for solution</summary>

```{r}

px %>% filter(skill == "Serve") %>%
    group_by(setter_position) %>%
    summarize(breakpoint_rate = mean(point_won_by == team))

```

</details>


<div id="footer"></div>