SorrTask830_PEOE_first_step (EDA)

research_amp/sports_prediciton/cricket_eda.py

@@ -0,0 +1,295 @@
+# ---
+# jupyter:
+#   jupytext:
+#     text_representation:
+#       extension: .py
+#       format_name: percent
+#       format_version: '1.3'
+#       jupytext_version: 1.15.2
+#   kernelspec:
+#     display_name: Python 3 (ipykernel)
+#     language: python
+#     name: python3
+# ---
+
+# %% [markdown]
+# # Datasets used:
+# https://www.kaggle.com/datasets/jamiewelsh2/ball-by-ball-ipl
+
+# %%
+import pandas as pd
+import numpy as np
+import matplotlib.pyplot as plt
+import seaborn as sns
+import matplotlib.pyplot as plt
+
+t20_data = pd.read_csv('datasets/cricket/ball_by_ball_ipl.csv')
+
+# %%
+print(t20_data.info())
+
+# %%
+t20_data['Bowler Runs Conceded'].value_counts()
+
+# %% [markdown]
+# ### Methods of Dismissal
+
+# %%
+# Filter data frame to get rid of N/A values in the method column.
+filtered_df = t20_data[t20_data['Method'] != 'N/A']
+# Count values in filtered data frame.
+value_counts = filtered_df['Method'].value_counts()
+ # Set the minimum occurrence count for displaying a slice.
+threshold = 50 
+# Filter out the methods with occurrence count below the threshold.
+filtered_value_counts = value_counts[value_counts >= threshold]
+# Create a custom "Other" category to group rare methods below the threshold.
+other_count = value_counts[value_counts < threshold].sum()
+filtered_value_counts['other'] = other_count
+# Capitalize the first letter of each label using str.capitalize().
+labels = filtered_value_counts.index.str.capitalize()
+# Plot the pie chart with the filtered value counts.
+plt.pie(filtered_value_counts, labels=labels, autopct='%.1f%%')
+plt.axis('equal')  # Equal aspect ratio ensures that pie is drawn as a circle.
+plt.title('Methods of Dismissal in International T20 Matches')
+plt.show()
+
+# %% [markdown]
+# ### Runs and Wickets per Over
+
+# %%
+overs = list(np.arange(1,21))
+Rpo = []
+Wpo = []
+for over in overs:
+    over_df = t20_data.loc[t20_data['Over'] == over]   
+    total_balls = len(over_df[over_df['Ball Rebowled'] == 0].index)
+    avg_wick_pb = over_df['Wicket'].sum()/(total_balls)
+    avg_pb = over_df['Runs From Ball'].sum()/(total_balls)
+    Wpo.append(avg_wick_pb*6)
+    Rpo.append(avg_pb*6)
+# Create a figure and axis object.
+fig , ax = plt.subplots(figsize=(15, 8))
+# Set the x-axis label.
+ax.set_xlabel('Over')
+# Set the y-axis label for the first field.
+ax.set_ylabel('Runs Per Over', color='k')
+# Create the bar chart for the first field.
+ax.bar(np.array(overs) - 0.15, Rpo, color='tab:purple', label='Runs per Over', width=0.3)
+# Create a second y-axis object.
+ax2 = ax.twinx()
+# Set the y-axis label for the second field.
+ax2.set_ylabel('Wickets Per Over', color='k')
+# Create the bar chart for the second field.
+ax2.bar(np.array(overs) + 0.15, Wpo, color='tab:cyan',label = 'Wickets per Over', width=0.3)
+# Set the x-axis limits to ensure all bars are visible.
+ax.set_xlim([0, 21])
+# create the legend.
+lines, labels = ax.get_legend_handles_labels()
+lines2, labels2 = ax2.get_legend_handles_labels()
+ax2.legend(lines + lines2, labels + labels2, loc='upper left')
+plt.xticks(np.arange(1, 20, step=1))
+# Show the plot.
+plt.title('Average Runs and Wickets per Over')
+plt.show()
+
+# %%
+years = ['2013', '2014', '2015', '2016', '2017', '2018', '2019', '2020', '2021', '2022', '2023']
+year_targets = []
+for year in years:
+    targets = []
+    year_df = t20_data[t20_data['Date'].str.slice(stop=4) == year]
+    for j in range(len(year_df.index)-1):
+        if year_df['Match ID'].iloc[j] != year_df['Match ID'].iloc[j+1]:
+            targets.append(year_df['Target Score'].iloc[j])
+    year_targets.append(sum(targets)/len(targets))
+df = pd.DataFrame({'Average Target Score': year_targets}, index=years)
+# Plot the data against years.
+df.plot(kind='line', color='purple')
+# Set the title and axis labels.
+plt.title('Average Target Score by Year')
+plt.xlabel('Year')
+plt.ylabel('Average Target Score')
+# Show the plot.
+plt.show()
+
+# %%
+powerplay_df = t20_data.loc[(t20_data['Over'] >= 1) & (t20_data['Over'] <= 6)]
+middle_df = t20_data.loc[(t20_data['Over'] >= 7) & (t20_data['Over'] <= 15)]
+end_df = t20_data.loc[(t20_data['Over'] >= 16) & (t20_data['Over'] <= 20)]
+
+player = 'JC Buttler'
+dfs =[powerplay_df, middle_df, end_df]
+AVGs = []
+SRs = []
+BASRAs = []
+for df in dfs:
+    player_mask = (df['Batter'] == player)
+    batter_df = df.loc[player_mask]
+    n_outs = len(df[df['Player Out'] == player].index)
+    total_runs = sum(batter_df['Batter Runs'])
+    total_bf = len(batter_df[batter_df['Ball Rebowled'] == 0].index)
+    bat_avg = total_runs/n_outs
+    sr = (total_runs/total_bf)*100
+    basra = bat_avg + sr
+    AVGs.append(bat_avg)
+    SRs.append(sr)
+    BASRAs.append(basra)
+# Example data.
+stages = ['Powerplay (1-6)', 'Middle Overs (7-15)', 'Final Overs (16-20)']
+# Set up the plot.
+fig, ax = plt.subplots(figsize=(8, 6))
+# Set the x-axis labels and positions.
+x_labels = stages
+x_positions = np.arange(len(x_labels))
+ax.bar(x_positions, SRs, width=0.4, label='Strike Rate', color='purple')
+# Plot the batting average bars.
+ax.bar(x_positions+0.4, AVGs, width=0.4, label='Batting Average', color = 'black')
+# Set the y-axis limits and labels.
+ax.set_ylim([0, 225])
+ax.set_ylabel('Strike Rate / Batting Average')
+# Set the title and legend.
+ax.set_title('Batting Performance by Phase of Play: JC Buttler')
+ax.legend()
+# Add the x-axis labels and tick marks.
+ax.set_xticks(x_positions+0.2)
+ax.set_xticklabels(x_labels)
+plt.show()
+
+# %%
+player = 'JC Buttler'
+
+# Slice the data frame by when player is at the crease.
+player_df =t20_data[(t20_data['Batter' ]== player) | (t20_data['Non Striker'] == player)]
+# Get a list of unique dates in date_col.
+unique_dates = player_df['Date'].unique()
+# Sort the unique dates in ascending order.
+unique_dates_sorted = sorted(unique_dates)
+player_data = []
+n_outs = 0
+total_runs = 0
+total_bf = 0
+for date in unique_dates_sorted:
+    game_df = player_df[player_df['Date'] == date]
+    batter_df = game_df[game_df['Batter'] == player]
+    runs = batter_df['Batter Runs'].sum()
+    # Check if any value in 'Player Out' column is equal to player.
+    if player in game_df['Player Out'].values:
+        n_outs += 1
+    bf = len(batter_df[batter_df['Ball Rebowled'] == 0].index)
+    total_runs += runs
+    total_bf += bf
+    player_data.append([player, date, total_runs/n_outs, (total_runs/total_bf)*100])
+player_data_df = pd.DataFrame(player_data, columns= ['Player', 'Date', 'Batting Average', 'Strike Rate'])
+# Set up the plot.
+fig, ax1 = plt.subplots()
+# Plot the first data set on the left y-axis.
+color = 'tab:red'
+ax1.set_xlabel('Date')
+ax1.set_ylabel('Batting Average', color=color)
+ax1.plot(player_data_df['Date'], player_data_df['Batting Average'], color=color)
+ax1.tick_params(axis='y', labelcolor=color)
+# Create a second y-axis on the right side of the plot.
+ax2 = ax1.twinx()
+# Plot the second data set on the right y-axis.
+color = 'k'
+ax2.set_ylabel('Strike Rate', color=color)
+ax2.plot(player_data_df['Date'], player_data_df['Strike Rate'], color=color)
+ax2.tick_params(axis='y', labelcolor=color)
+# Set the tick frequency on the x-axis.
+ax1.xaxis.set_major_locator(plt.MaxNLocator(6))
+# Display legend.
+ax.legend()
+# Format the dates on the x-axis.
+fig.autofmt_xdate()
+plt.title('JC Buttler Running Batting Statistics')
+# Show the plot.
+plt.show()
+
+# %%
+# Convert the "date" column to datetime format.
+middle_df['Date'] = pd.to_datetime(middle_df['Date'])
+# Slice the dataframe by values in the "date" column after the year 2015.
+middle_df = middle_df[middle_df['Date'].dt.year > 2015]
+# Get list of batters that have been at the crease during the middle overs in this period.
+batter_list = middle_df['Batter'].unique().tolist()
+BATTER = []
+BASRA = []
+BASRA2 = []
+for batter in batter_list:
+    batter_df = middle_df.loc[middle_df['Batter'] == batter]  
+    if len(batter_df.index) > 200:
+        BATTER.append(batter)
+        n_outs = len(middle_df[middle_df['Player Out'] == batter].index)      
+        total_runs = sum(batter_df['Batter Runs'])
+        total_bf = len(batter_df[batter_df['Ball Rebowled'] == 0].index)
+        bat_avg = total_runs/n_outs
+        sr = (total_runs/total_bf)*100
+        BASRA.append(bat_avg + sr)
+        BASRA2.append(2*bat_avg + sr)
+# convert the list to a numpy array.
+BASRA = np.array(BASRA)
+# get the indices of the 8 max values.
+max_indices = np.argsort(BASRA)[-8:]
+max_batters = [BATTER[i]for i in max_indices]
+max_BASRAs = [BASRA[i] for i in max_indices]
+fig, ax = plt.subplots(figsize=(15, 8))
+x_labels = max_batters
+x_positions = np.arange(len(x_labels))
+ax.bar(x_positions, max_BASRAs, width=0.4, label='BASRA', color='purple')\
+# Set the y-axis limits and labels.
+ax.set_ylim([0, 300])
+ax.set_ylabel('BASRA')
+# Add the x-axis labels and tick marks.
+ax.set_xticks(x_positions)
+ax.set_xticklabels(x_labels)
+# Set the title and legend.
+ax.set_title('Highest Performing Batters in Middle Overs since 2016')
+plt.show()
+
+# %%
+t20_data['Date'] = pd.to_datetime(t20_data['Date'])
+starts = t20_data[t20_data['Date'].dt.year > 2015]
+starts.loc[:, 'Valid Ball'] = (starts['Ball Rebowled'] != 1).astype(int)
+batter_list = starts['Batter'].unique().tolist()
+SRs = []
+batters = []
+for batter in batter_list:
+
+    batter_df = starts.loc[starts['Batter'] == batter]
+    ids = batter_df['Match ID'].unique()
+    first10 = []
+    for match_id in ids:
+        game_df = batter_df[batter_df['Match ID'] == match_id]
+        cum_sum = 0
+        index_to_stop = 0
+        for index, row in game_df.iterrows():
+            cum_sum += row['Valid Ball']
+            if cum_sum >= 10:
+                index_to_stop = index + 1
+                break
+        game_df = game_df.iloc[:index_to_stop]
+        first10.append(game_df)
+    first10 = pd.concat(first10, axis=0)
+    if len(first10.index) > 100:
+        batters.append(batter)
+        SRs.append((first10['Batter Runs'].sum()/len(first10[first10['Ball Rebowled'] == 0].index))*100)
+# convert the list to a numpy array.
+SRs = np.array(SRs)
+# get the indices of the 8 max values.
+max_indices = np.argsort(SRs)[-8:]
+max_batters = [batters[i]for i in max_indices]
+max_SRs = [SRs[i] for i in max_indices]
+fig, ax = plt.subplots(figsize=(15, 8))
+x_labels = max_batters
+x_positions = np.arange(len(x_labels))
+ax.bar(x_positions, max_SRs, width=0.4, label='First 10 Balls of Innings Strike Rate', color='purple')
+# Set the y-axis limits and labels.
+ax.set_ylim([0, 300])
+ax.set_ylabel('Strike Rate')
+# Add the x-axis labels and tick marks.
+ax.set_xticks(x_positions)
+ax.set_xticklabels(x_labels)
+# Set the title and legend.
+ax.set_title('Fastest Starting Batters in IT20s since 2016')
+plt.show()