webcast_coverage.Rmd

---
title: "TBA Webcast Coverage"
author: "Greg Marra"
date: "11/4/2018"
output: github_document
---

Many people come to The Blue Alliance to watch webcasts of FIRST Robotics Competition events. How is our coverage ofwebcasts and match videos trending over time?

# Webcasts

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = FALSE)

library(tidyverse)
library(skimr)

source("get_tba_data.R")

get_size <- function(thing) {
  if ( class(thing) == "list" ) {
    return(thing %>% length)
  } else if ( class(thing) == "data.frame") {
    return(thing %>% nrow)
  } else {
    return(FALSE)
  }
}

years <- c(
  2011,
  2012,
  2013,
  2014,
  2015,
  2016, 
  2017, 
  2018)

events <- years %>%
  map(~ getEvents(.x)) %>%
  bind_rows() %>%
  mutate(
    has_webcast = map_int(webcasts, get_size) > 0,
    official = !event_type_string %in% c("Preseason", "Offseason")
  ) %>%
  mutate(
    official = factor(.$official, levels=c(TRUE, FALSE)) # put official events before unofficial ones
  )

matches <- map2(events$year, 
                events$event_code, 
                ~ getEventMatches(.x, .y)
                ) %>%
  bind_rows()

matches <- matches %>%
  left_join(events, by = c("event_key" = "key")) %>%
  mutate(
    has_video = map_int(videos, get_size) > 0
  )


```

### Event Webcast Coverage

```{r webcast_coverage_bar_chart}
events %>%
  ggplot(aes(x = year,
             fill = has_webcast)) +
  geom_bar() +
  facet_grid(. ~ official, labeller = label_both) +
  labs(
    title = "Event Webcast Coverage",
    subtitle = "Webcast coverage is growing over time",
    x = "Year",
    y = "Events",
    fill = "Has Webcast"
  )
```

### As Ratios

While nearly all official events are webcast, we see that the trend over time is for more and more unofficial events to also run webcasts.

```{r webcast_coverage_pct}
events %>%
  group_by(year, official) %>%
  summarize(
    n_events = n(),
    n_events_w_webcast = sum(has_webcast),
    pct_events_w_webcast = mean(has_webcast)
  ) %>%
  ggplot(aes(x = year,
             y = pct_events_w_webcast,
             color = official)) +
  geom_point() +
  labs(
    title = "Event Webcast Coverage",
    subtitle = "Official events all have webcasts, and offseasons are growing in coverage",
    x = "Year",
    y = "Event Webcast Coverage",
    color = "Official Event?"
  )
```

## Match Video Coverage

```{r match_video_coverage}
# select(-starts_with("score_breakdown")) %>%

matches %>%
  ggplot(aes(x = year,
             fill = has_video)) +
  geom_bar() +
  facet_grid(. ~ official, labeller = label_both) +
  labs(
    title = "Match Video Coverage",
    subtitle = "Video coverage is growing over time, but many offseasons events don't have match data at all.",
    x = "Year",
    y = "Matches",
    fill = "Has Videos"
  )

matches %>%
  group_by(year, official) %>%
  summarize(
    n_matches = n(),
    n_matches_w_video = sum(has_video),
    pct_matches_w_video = mean(has_video)
  ) %>%
  ggplot(aes(x = year,
             y = pct_matches_w_video,
             color = official)) +
  geom_point() +
  labs(
    title = "Match Video Coverage",
    subtitle = "Match video coverage is growing over time for official events, but shrinking for unofficial events",
    x = "Year",
    y = "Match Video Coverage",
    color = "At Official Event?"
  )
```