Packages

This section ensures that all required libraries for data extraction, manipulation, and visualization are installed and loaded. It combines general-purpose data science tools with football-specific analytics packages.

# Install packages
if (!require("devtools")) install.packages("devtools")
if (!require("soccermatics")) devtools::install_github("jogall/soccermatics")

# Load packages
library(soccermatics)
library(tidyverse)
library(ggplot2)
library(StatsBombR)

Data loading

This section retrieves match data from the StatsBomb open dataset, filters for a specific competition and match, and prepares the data for analysis by cleaning and transforming spatial coordinates.

# Show all competitions available with Statsbomb
comps <- FreeCompetitions()

# Choose one competition
competition <- comps %>% 
  filter(competition_name == "Premier League", 
         season_name == "2015/2016")

# Load matches from the selected competition
matches_available <- FreeMatches(competition)

# Choose one match
match <- matches_available %>% filter(
  home_team.home_team_name == "Manchester City",
  away_team.away_team_name == "Southampton")

# Load events from the selected match
events <- free_allevents(MatchesDF = match, Parallel = TRUE)

# Clean the data (remove unnecessary columns, handle missing values)
events <- allclean(events)

# Transform coordinates to the pitch dimensions required by Soccermatics
clean_events <- events %>%
  soccerTransform(method='statsbomb')

Analysis

1. Offensive Analysis

a - Shot Map with xG values

This section evaluates the offensive performance of both teams using shot maps enriched with expected goals (xG), allowing both spatial and probabilistic assessment of chance quality.

Manchester City

clean_events %>% 
  filter(team.name == "Manchester City") %>%
  soccerShotmap(
    theme = "dark",
    title = "Shots by Manchester City with xG",
    subtitle = "Manchester City 3-1 Southampton | Premier League game (2015/2016)"
  )

Interpretation:
- Manchester City generated 3 goals from 1.67 xG, indicating efficient finishing above expectation.
- The shot distribution shows multiple attempts in and around the penalty area, with at least one high-quality chance (large xG value near goal).
- The spatial pattern suggests central penetration and effective chance creation in dangerous zones.

Southampton

clean_events %>% 
  filter(team.name == "Southampton") %>%
  soccerShotmap(
    theme = "dark",
    title = "Shots by Southampton with xG",
    subtitle = "Manchester City 3-1 Southampton | Premier League game (2015/2016)"
  )

Interpretation:
- Southampton scored 1 goal from 1.16 xG, which is broadly in line with expectation.
- Their shots are more dispersed and slightly further from goal, indicating lower average shot quality.
- Fewer high-probability chances suggest difficulty in breaking into optimal shooting areas.

b - xG timeline

The xG timeline tracks the cumulative expected goals throughout the match, providing a dynamic view of how scoring opportunities evolved over time.

clean_events %>% 
  soccerxGTimeline(
    homeCol = "#66615e", 
    awayCol = "#a50044", 
    labels = FALSE
  )

Interpretation:
- Manchester City shows a rapid early increase in xG, suggesting strong early attacking pressure.
- Their xG curve grows steadily, indicating consistent chance creation across the match.
- Southampton’s xG increases more gradually, with a noticeable rise later in the match, implying delayed offensive momentum.
- The final totals (~1.7 vs ~1.2) confirm a moderate but clear advantage for Manchester City.

2. Passing Analysis

a - Passing Network

Passing networks illustrate team structure by mapping player positions and pass connections, highlighting distribution patterns and key players in build-up play.

Manchester City

clean_events %>%
  filter(team.name == "Manchester City") %>%
  soccerPassmap(
    fill = "#66615e",
    arrow = "r",
    shortName = FALSE, 
    title = "Manchester City Pass Network | Premier League game vs Southampton (2015/2016)"
  )

Interpretation:
- The network is dense and well-connected, particularly in midfield.
- Players like Yaya Touré and Kevin De Bruyne appear central, acting as key hubs.
- The structure indicates possession-based play with short passing combinations and strong central control.
- High pass completion (77.6%) reinforces technical dominance and control.

Southampton

clean_events %>%
  filter(team.name == "Southampton") %>%
  soccerPassmap(
    fill = "#66615e",
    arrow = "r",
    shortName = FALSE, 
    title = "Southampton Pass Network | Premier League game vs Man City (2015/2016)"
  )

Interpretation:
- The pass network appears to be generated over a restricted time window by the plotting function, likely due to its internal handling of substitutions or lineup changes.

b - Passing map by player

This section zooms into individual player behavior, showing pass direction, length, and type to better understand player roles within the system.

Manchester City - Kevin de Bruyne

clean_events %>%
  filter(team.name == "Manchester City" & 
           type.name == "Pass" & 
           player.name == "Kevin De Bruyne") %>%
  {
    soccerPitch(arrow = "r",
                title = "Kevin De Bruyne Pass Map vs Southampton",
                subtitle = "Manchester City 3-1 Southampton | Premier League game (2015/2016)") +
    geom_segment(data = ., aes(x = location.x, y = location.y,
                               xend = pass.end_location.x, yend = pass.end_location.y,
                               color = pass.height.name),
                 arrow = arrow(length = unit(0.25, "cm"), type = "closed")) +
    theme(legend.position = "bottom",
          plot.margin = margin(5,5,5,5),
          legend.margin = margin(0,0,10,0)) +
    labs(color = "Pass Height")
  }

Interpretation:
- Highly active across the attacking half, with forward-oriented and varied passes.
- Presence of long and high passes indicates creative playmaking and vertical progression.
- His distribution pattern confirms his role as a primary offensive orchestrator.

Southampton - Oriol Romeu Vidal

clean_events %>%
  filter(team.name == "Southampton" & 
           type.name == "Pass" & 
           player.name == "Oriol Romeu Vidal") %>%
  {
    soccerPitch(arrow = "r",
                title = "Oriol Romeu Vidal Pass Map vs Manchester City",
                subtitle = "Manchester City 3-1 Southampton | Premier League game (2015/2016)") +
    geom_segment(data = ., aes(x = location.x, y = location.y,
                               xend = pass.end_location.x, yend = pass.end_location.y,
                               color = pass.height.name),
                 arrow = arrow(length = unit(0.25, "cm"), type = "closed")) +
    theme(legend.position = "bottom",
          plot.margin = margin(5,5,5,5),
          legend.margin = margin(0,0,10,0)) +
    labs(color = "Pass Height")
  }

Interpretation:
- Passes are more central and conservative, with fewer progressive actions.
- Distribution is largely lateral or short-range, indicating a holding midfield role.
- Limited penetration suggests support rather than creation.

c - Passing heatmap and flow

This section combines pass density (heatmap) and directional flow to visualize where and how teams move the ball across the pitch.

Manchester City

# Standardize columns to allow layering multiple visualizations
clean_events %>%
  filter(team.name == "Manchester City" & type.name == "Pass") %>%
  soccerStandardiseCols(method = 'statsbomb') -> mancity_pass_data

# Heatmap of pass density
heatmap_plot <- soccerHeatmap(mancity_pass_data, 
                              xBins = 6, yBins = 6,
                              arrow = "r",
                              colLow = "#fff9cc", colHigh = "#ff6100",
                              title = "Manchester City Pass Flow",
                              subtitle = "Manchester City 3-1 Southampton | Premier League game (2015/2016)")

# Overlay flow of passes
soccerFlow(mancity_pass_data,
           xBins = 6, yBins = 6,
           plot = heatmap_plot, # superimpose with heatmap
           col = "#37577d",
           lwd = 0.8)

Interpretation:
- High pass density in central and advanced midfield zones.
- Flow vectors indicate forward progression with structured buildup.
- Strong presence in the opponent’s half reflects territorial dominance.

Southampton

# Standardize columns to allow layering multiple visualizations
clean_events %>%
  filter(team.name == "Southampton" & type.name == "Pass") %>%
  soccerStandardiseCols(method = 'statsbomb') -> southampton_pass_data

# Heatmap of pass density
heatmap_plot <- soccerHeatmap(southampton_pass_data, 
                              xBins = 6, yBins = 6,
                              arrow = "r",
                              colLow = "#fff9cc", colHigh = "#ff6100",
                              title = "Southampton Pass Flow",
                              subtitle = "Manchester City 3-1 Southampton | Premier League game (2015/2016)")

# Overlay flow of passes
soccerFlow(southampton_pass_data,
           xBins = 6, yBins = 6,
           plot = heatmap_plot, # superimpose with heatmap
           col = "#37577d",
           lwd = 0.8)

Interpretation:
- Heatmap shows concentration in midfield areas, especially centrally.
- Less activity in the final third suggests difficulty progressing into dangerous zones.
- Flow appears more horizontal and less penetrative.

3. Pressing Analysis

a - Heatmap

This section analyzes defensive intensity through pressure events, highlighting where teams attempt to regain possession.

Manchester City

# Identify events with valid coordinates
position_events <- clean_events %>%
  filter(!is.na(location.x)) %>%
  select(type.name) %>%
  unique() %>%
  as.list()

# Different types of events
#position_events$type.name

# Heatmap of all actions by Kevin De Bruyne (Manchester City)
clean_events %>%
  filter(team.name == "Manchester City" & player.name == "Kevin De Bruyne" &
           type.name %in% position_events$type.name) %>%
  soccerHeatmap(
    x = "location.x", y = "location.y",
    kde = TRUE,          # smoothed heatmap
    arrow = "r",         # direction of attack
    title = "Kevin De Bruyne Heatmap",
    subtitle = "Manchester City 3-1 Southampton | Premier League game (2015/2016)"
  )

# Heatmap of pressure zones for Manchester City
clean_events %>%
  filter(team.name == "Manchester City" & type.name == "Pressure") %>%
  soccerHeatmap(
    x = "location.x", y = "location.y",
    xBins = 6, yBins = 3,   # grid for counting events
    colLow = "#fff9cc", colHigh = "#ff6100",
    arrow = "r",
    title = "Manchester City Pressure Zones",
    subtitle = "Manchester City 3-1 Southampton | Premier League game (2015/2016)"
  )

Interpretation:

Kevin De Bruyne (Heatmap) :
- High activity in the attacking midfield zone, indicating high pressing involvement.
- Suggests Manchester City applies pressure high up the pitch.
Manchester City (Pressure Zones) :
- Concentration of pressure in advanced and midfield areas.
- Indicates a high pressing system, aiming to recover the ball early.

Southampton

# First, identify events with valid coordinates
position_events <- clean_events %>%
  filter(!is.na(location.x)) %>%
  select(type.name) %>%
  unique() %>%
  as.list()

# different types of events
position_events$type.name

##  [1] "Pass"           "Ball Receipt*"  "Carry"          "Pressure"      
##  [5] "Dispossessed"   "Duel"           "Foul Committed" "Foul Won"      
##  [9] "Interception"   "Shot"           "Goal Keeper"    "Clearance"     
## [13] "Ball Recovery"  "Block"          "Miscontrol"     "Dribbled Past" 
## [17] "Dribble"        "Offside"

# Heatmap of all actions by Oriol Romeu Vidal (Southampton)
clean_events %>%
  filter(team.name == "Southampton" & player.name == "Oriol Romeu Vidal" &
           type.name %in% position_events$type.name) %>%
  soccerHeatmap(
    x = "location.x", y = "location.y",
    kde = TRUE,          # smoothed heatmap
    arrow = "r",         # direction of attack
    title = "Oriol Romeu Vidal Heatmap",
    subtitle = "Manchester City 3-1 Southampton | Premier League game (2015/2016)"
  )

# Heatmap of pressure zones for Southampton
clean_events %>%
  filter(team.name == "Southampton" & type.name == "Pressure") %>%
  soccerHeatmap(
    x = "location.x", y = "location.y",
    xBins = 6, yBins = 3,   # grid for counting events
    colLow = "#fff9cc", colHigh = "#ff6100",
    arrow = "r",
    title = "Southampton Pressure Zones",
    subtitle = "Manchester City 3-1 Southampton | Premier League game (2015/2016)"
  )

Interpretation:

Oriol Romeu (Heatmap) :
- Activity concentrated in central defensive midfield zones.
- Reflects a deeper, more positional defensive role.
Southampton (Pressure Zones) :
- Pressure occurs mostly in midfield and defensive thirds.
- Indicates a mid-to-low block defensive approach.

Collaborative Task - Match Analysis through Offensive, Passing and Pressing aspects

Pierre Anorga

2026-04-21

Packages

Data loading

Analysis

1. Offensive Analysis

a - Shot Map with xG values

Manchester City

Southampton

b - xG timeline

2. Passing Analysis

a - Passing Network

Manchester City

Southampton

b - Passing map by player

Manchester City - Kevin de Bruyne

Southampton - Oriol Romeu Vidal

c - Passing heatmap and flow

Manchester City

Southampton

3. Pressing Analysis

a - Heatmap

Manchester City

Southampton