fix plot

src/02 Exploratory data analysis.ipynb

@@ -170,9 +170,47 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 4,
+   "execution_count": 10,
    "metadata": {},
    "outputs": [],
+   "source": [
+    "head_coach = pd.read_csv('data/head_coach.csv', dtype = {'HeadCoach' : 'str'}, parse_dates=['Appointed', 'EndDate'])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Seasons: 2014-2023\n",
+      "Leagues: Premier League, La Liga, Ligue 1, Bundesliga, Serie A\n",
+      "Number of matches: 14873\n",
+      "Number of teams: 161\n",
+      "Number of teams without coach: 96\n",
+      "Percentage of matches without coach: 44.97%\n",
+      "Number of appointments: 398\n",
+      "Number of unique coaches: 233\n",
+      "Number of teams per coach:              HeadCoach  n_team\n",
+      "178     Rafael Benítez       4\n",
+      "142     Maurizio Sarri       4\n",
+      "223         Unai Emery       4\n",
+      "176      Quique Setién       4\n",
+      "81    Giuseppe Iachini       4\n",
+      "..                 ...     ...\n",
+      "93         Javi Gracia       1\n",
+      "94      Javier Calleja       1\n",
+      "95   Jean-Louis Gasset       1\n",
+      "96    Jean-Luc Vasseur       1\n",
+      "116     Louis van Gaal       1\n",
+      "\n",
+      "[233 rows x 2 columns]\n"
+     ]
+    }
+   ],
    "source": [
     "min_season = match_results['Date'].dt.year.min()\n",
     "max_season = match_results['Date'].dt.year.max()\n",
@@ -182,7 +220,18 @@
     "n_team_no_coach = match_results.groupby(['Team'])['HeadCoach'].apply(lambda x: x.isnull().all()).sum()\n",
     "perc_match_no_coach = match_results['HeadCoach'].isnull().sum() / len(match_results) * 100\n",
     "n_unique_coach = match_results['HeadCoach'].nunique()\n",
-    "n_unique_coach_records = match_results.groupby(['HeadCoach'])['Team'].nunique().reset_index(name='n_team').sort_values(by='n_team', ascending=False)"
+    "n_appointments = len(head_coach)\n",
+    "n_unique_coach_records = match_results.groupby(['HeadCoach'])['Team'].nunique().reset_index(name='n_team').sort_values(by='n_team', ascending=False)\n",
+    "\n",
+    "print(f\"Seasons: {min_season}-{max_season}\")\n",
+    "print(f\"Leagues: {leagues}\")\n",
+    "print(f\"Number of matches: {n_match}\")\n",
+    "print(f\"Number of teams: {n_team}\")\n",
+    "print(f\"Number of teams without coach: {n_team_no_coach}\")\n",
+    "print(f\"Percentage of matches without coach: {perc_match_no_coach:.2f}%\")\n",
+    "print(f\"Number of appointments: {n_appointments}\")\n",
+    "print(f\"Number of unique coaches: {n_unique_coach}\")\n",
+    "print(f\"Number of teams per coach: {n_unique_coach_records}\")"
    ]
   },
   {
@@ -351,18 +400,40 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 7,
+   "execution_count": 13,
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Average home goals: 1.51\n",
+      "Average away goals: 1.18\n",
+      "Difference in average goals: -21.52%\n",
+      "Home win percentage: 0.43\n",
+      "Away win percentage: 0.29\n",
+      "Difference in win percentage: -48.77%\n",
+      "Draw percentage: 0.26\n"
+     ]
+    }
+   ],
    "source": [
     "home_goal = match_results[match_results['isHome']]['Goals'].mean()\n",
     "away_goal = match_results[~match_results['isHome']]['Goals'].mean()\n",
     "diff_goal_perc = ((away_goal - home_goal) / home_goal) * 100\n",
     "\n",
     "home_win = match_results[match_results['isHome']]['Result'].value_counts(normalize=True)['win']\n",
     "away_win = match_results[~match_results['isHome']]['Result'].value_counts(normalize=True)['win']\n",
-    "diff_win_perc = ((away_win - home_win) / home_win) * 100\n",
-    "home_draw = match_results[match_results['isHome']]['Result'].value_counts(normalize=True)['draw']"
+    "diff_win_perc = ((away_win - home_win) / away_win) * 100\n",
+    "home_draw = match_results[match_results['isHome']]['Result'].value_counts(normalize=True)['draw']\n",
+    "\n",
+    "print(f\"Average home goals: {home_goal:.2f}\")\n",
+    "print(f\"Average away goals: {away_goal:.2f}\")\n",
+    "print(f\"Difference in average goals: {diff_goal_perc:.2f}%\")\n",
+    "print(f\"Home win percentage: {home_win:.2f}\")\n",
+    "print(f\"Away win percentage: {away_win:.2f}\")\n",
+    "print(f\"Difference in win percentage: {diff_win_perc:.2f}%\")\n",
+    "print(f\"Draw percentage: {home_draw:.2f}\")"
    ]
   },
   {