From d92f4f8161ef8f65e966dc12062342fe18b07a26 Mon Sep 17 00:00:00 2001 From: "osboxes.org" Date: Sun, 27 Mar 2022 07:37:55 -0400 Subject: [PATCH] dodanie ium03 --- IUM03.ipynb | 543 ++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 543 insertions(+) create mode 100644 IUM03.ipynb diff --git a/IUM03.ipynb b/IUM03.ipynb new file mode 100644 index 0000000..86a586f --- /dev/null +++ b/IUM03.ipynb @@ -0,0 +1,543 @@ +{ + "cells": [ + { + "cell_type": "markdown", + "id": "7100de70", + "metadata": {}, + "source": [ + "1. Pobieranie bazy" + ] + }, + { + "cell_type": "code", + "execution_count": 6, + "id": "17518731", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Warning: Your Kaggle API key is readable by other users on this system! To fix this, you can run 'chmod 600 /home/osboxes/.kaggle/kaggle.json'\n", + "Downloading extended-football-stats-for-european-leagues-xg.zip to /home/osboxes/jupyter_dir/notebooks/IUM03\n", + " 73%|███████████████████████████▋ | 1.00M/1.37M [00:00<00:00, 5.12MB/s]\n", + "100%|██████████████████████████████████████| 1.37M/1.37M [00:00<00:00, 3.95MB/s]\n" + ] + } + ], + "source": [ + "!kaggle datasets download -d slehkyi/extended-football-stats-for-european-leagues-xg" + ] + }, + { + "cell_type": "code", + "execution_count": 7, + "id": "0d9abb90", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Archive: extended-football-stats-for-european-leagues-xg.zip\n", + " inflating: understat.com.csv \n", + " inflating: understat_per_game.csv \n" + ] + } + ], + "source": [ + "!unzip -o extended-football-stats-for-european-leagues-xg.zip" + ] + }, + { + "cell_type": "markdown", + "id": "edcb45ce", + "metadata": {}, + "source": [ + "2. Zmiana nazwy plikow" + ] + }, + { + "cell_type": "code", + "execution_count": 8, + "id": "cb6595a3", + "metadata": {}, + "outputs": [], + "source": [ + "mv understat.com.csv understat.csv" + ] + }, + { + "cell_type": "markdown", + "id": "b4aea8ec", + "metadata": {}, + "source": [ + "3. Zmiana nazwy kolumn" + ] + }, + { + "cell_type": "code", + "execution_count": 9, + "id": "6c519048", + "metadata": {}, + "outputs": [], + "source": [ + "import pandas as pd\n", + "understat = pd.read_csv('understat.csv')\n", + "understat_per_game = pd.read_csv('understat_per_game.csv')" + ] + }, + { + "cell_type": "code", + "execution_count": 18, + "id": "fc730d10", + "metadata": {}, + "outputs": [], + "source": [ + "understat.rename( columns={'Unnamed: 0':'league'}, inplace=True)\n", + "understat.rename( columns={'Unnamed: 1':'year'}, inplace=True)" + ] + }, + { + "cell_type": "code", + "execution_count": 19, + "id": "5b636526", + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
leagueyearpositionteammatcheswinsdrawslosesscoredmissed...xGAxGA_diffnpxGAnpxGDppda_coefoppda_coefdeepdeep_allowedxptsxpts_diff
0La_liga20141Barcelona38304411021...28.4442937.44429324.72790773.0493055.68353516.36759348911494.08130.0813
1La_liga20142Real Madrid38302611838...42.6071984.60719838.89080547.21309010.20908512.92951035115381.7489-10.2511
2La_liga20143Atletico Madrid3823966729...29.0691070.06910726.83927125.7487378.9820289.23709119712373.1353-4.8647
3La_liga20144Valencia38221157032...39.3925727.39257233.44647716.2575018.7098277.87022520317263.7068-13.2932
4La_liga20145Sevilla3823787145...47.8627422.86274241.91652920.1780708.2761489.47780530516867.3867-8.6133
\n", + "

5 rows × 24 columns

\n", + "
" + ], + "text/plain": [ + " league year position team matches wins draws loses \\\n", + "0 La_liga 2014 1 Barcelona 38 30 4 4 \n", + "1 La_liga 2014 2 Real Madrid 38 30 2 6 \n", + "2 La_liga 2014 3 Atletico Madrid 38 23 9 6 \n", + "3 La_liga 2014 4 Valencia 38 22 11 5 \n", + "4 La_liga 2014 5 Sevilla 38 23 7 8 \n", + "\n", + " scored missed ... xGA xGA_diff npxGA npxGD ppda_coef \\\n", + "0 110 21 ... 28.444293 7.444293 24.727907 73.049305 5.683535 \n", + "1 118 38 ... 42.607198 4.607198 38.890805 47.213090 10.209085 \n", + "2 67 29 ... 29.069107 0.069107 26.839271 25.748737 8.982028 \n", + "3 70 32 ... 39.392572 7.392572 33.446477 16.257501 8.709827 \n", + "4 71 45 ... 47.862742 2.862742 41.916529 20.178070 8.276148 \n", + "\n", + " oppda_coef deep deep_allowed xpts xpts_diff \n", + "0 16.367593 489 114 94.0813 0.0813 \n", + "1 12.929510 351 153 81.7489 -10.2511 \n", + "2 9.237091 197 123 73.1353 -4.8647 \n", + "3 7.870225 203 172 63.7068 -13.2932 \n", + "4 9.477805 305 168 67.3867 -8.6133 \n", + "\n", + "[5 rows x 24 columns]" + ] + }, + "execution_count": 19, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "understat.head()" + ] + }, + { + "cell_type": "code", + "execution_count": 21, + "id": "cbf956d9", + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
leagueyearh_axGxGAnpxGnpxGAdeepdeep_allowedscored...ppda_coefppda_attppda_defoppda_coefoppda_attoppda_defteamxG_diffxGA_diffxpts_diff
0Bundesliga2014h2.570121.1984202.570121.198420542...9.6250002312421.85000043720Bayern Munich0.570120.198420-0.6514
1Bundesliga2014a1.503281.3079501.503281.3079501011...4.7560981954117.69565240723Bayern Munich0.503280.3079500.5143
2Bundesliga2014h1.229870.3101661.229870.3101661332...5.0606061673316.96153844126Bayern Munich-0.770130.310166-0.8412
3Bundesliga2014a1.035190.2031181.035190.203118620...4.423077115269.44680944447Bayern Munich1.035190.2031181.1367
4Bundesliga2014h3.482860.4028443.482860.4028442324...4.2500001704044.80000044810Bayern Munich-0.517140.402844-0.0713
\n", + "

5 rows × 29 columns

\n", + "
" + ], + "text/plain": [ + " league year h_a xG xGA npxG npxGA deep \\\n", + "0 Bundesliga 2014 h 2.57012 1.198420 2.57012 1.198420 5 \n", + "1 Bundesliga 2014 a 1.50328 1.307950 1.50328 1.307950 10 \n", + "2 Bundesliga 2014 h 1.22987 0.310166 1.22987 0.310166 13 \n", + "3 Bundesliga 2014 a 1.03519 0.203118 1.03519 0.203118 6 \n", + "4 Bundesliga 2014 h 3.48286 0.402844 3.48286 0.402844 23 \n", + "\n", + " deep_allowed scored ... ppda_coef ppda_att ppda_def oppda_coef \\\n", + "0 4 2 ... 9.625000 231 24 21.850000 \n", + "1 1 1 ... 4.756098 195 41 17.695652 \n", + "2 3 2 ... 5.060606 167 33 16.961538 \n", + "3 2 0 ... 4.423077 115 26 9.446809 \n", + "4 2 4 ... 4.250000 170 40 44.800000 \n", + "\n", + " oppda_att oppda_def team xG_diff xGA_diff xpts_diff \n", + "0 437 20 Bayern Munich 0.57012 0.198420 -0.6514 \n", + "1 407 23 Bayern Munich 0.50328 0.307950 0.5143 \n", + "2 441 26 Bayern Munich -0.77013 0.310166 -0.8412 \n", + "3 444 47 Bayern Munich 1.03519 0.203118 1.1367 \n", + "4 448 10 Bayern Munich -0.51714 0.402844 -0.0713 \n", + "\n", + "[5 rows x 29 columns]" + ] + }, + "execution_count": 21, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "understat_per_game.head()" + ] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 3 (ipykernel)", + "language": "python", + "name": "python3" + }, + "language_info": { + "codemirror_mode": { + "name": "ipython", + "version": 3 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython3", + "version": "3.8.10" + } + }, + "nbformat": 4, + "nbformat_minor": 5 +}