Махди Юсуф записал 20-ти минутный скринкаст с данными, которые мне удалось найти по имени файла Kevin Durant 2012-13 Game Log Он анализирует действия игрока Кевина Дюранта...
Файл csv пришлось предварительно отредактировать, создание новой таблицы, процедура замены строки "40:00" числом (секунд) -2400, группировка... Пост заканчивается диаграммой, построенной с помощью модуля vincent - Vega

При импорте vincent возникли трудности, так как не была задокументирована команда "from IPython.display import display, HTML, Javascript". Решение было найдено благодаря http://nbviewer.ipython.org/gist/anonymous/5436794

In [1]:

from IPython.display import YouTubeVideo
YouTubeVideo('BM7j6YGOv7U')

Out[1]:

Introductijn Pandas DataFrames¶

In [60]:

# do not forget to do it if it needs
from IPython.display import display, HTML, Javascript
# %pylab inline

In [3]:

import pandas as pd

In [7]:

pd.set_option('display.max_columns',None)

In [8]:

# I forgot ... I`ve already imported Pandas
from pandas import DataFrame, Series

In [61]:

# https://pypi.python.org/pypi/vincent
import vincent
vincent.core.initialize_notebook()

<IPython.core.display.Javascript at 0xb3fb780>

<IPython.core.display.Javascript at 0xb3fb780>

Importing csv data¶

As you can see the CSV file has two defects (... 'Tm','','Opp','','GS',...) - missing items in header row. And huge problem with rows (header row is repeating after every 20 rows). But there are not any remarks of it in video...
I have removed repeated header rows and insert two missing header strings literals 'H/A','W/L'

In [11]:

columns=['Rk','G','Date','Age','Tm','H/A','Opp','W/L','GS','MP','FG','FGA','FG%','3P','3PA','3P%','FT','FTA','FT%','ORB','DRB','TRB','AST','STL','BLK','TOV','PF','PTS','GmSc','+/-']

In [37]:

data=pd.read_csv('../data/KevinDuran2012-13GameLog.csv',names=columns)
data.head()

Out[37]:

	Rk	G	Date	Age	Tm	H/A	Opp	W/L	GS	MP	FG	FGA	FG%	3P	3PA	3P%	FT	FTA	FT%	ORB	DRB	TRB	AST	STL	BLK	TOV	PF	PTS	GmSc	+/-
0	1	1	2012-11-01	24-033	OKC	@	SAS	L (-2)	1	40:39	9	18	0.500	1	2	0.500	4	5	0.800	2	12	14	5	2	0	4	1	23	19.7	2
1	2	2	2012-11-02	24-034	OKC	NaN	POR	W (+14)	1	42:41	7	14	0.500	1	3	0.333	8	12	0.667	2	15	17	7	0	2	6	1	23	20.2	13
2	3	3	2012-11-04	24-036	OKC	NaN	ATL	L (-9)	1	42:05	7	17	0.412	1	4	0.250	7	8	0.875	1	11	12	8	3	2	6	3	22	19.3	-8
3	4	4	2012-11-06	24-038	OKC	NaN	TOR	W (+20)	1	29:09	4	11	0.364	1	4	0.250	6	6	1.000	0	6	6	3	2	0	4	3	15	9.6	23
4	5	5	2012-11-08	24-040	OKC	@	CHI	W (+6)	1	38:23	11	19	0.579	0	2	0.000	2	2	1.000	1	3	4	1	3	3	6	1	24	16.1	-3

5 rows × 30 columns

Here you can dig the explanations of all literals

In []:

...<tr class="">
  <th data-stat="ranker" align="right" class="ranker sort_default_asc show_partial_when_sorting" tip="Rank">Rk</th>
  <th data-stat="game_season" align="right" class="tooltip" tip="Season Game">G</th>
  <th data-stat="date_game" align="left" class="tooltip sort_default_asc">Date</th>
  <th data-stat="age" align="center" class="tooltip sort_default_asc" tip="Age of Player at the start of February 1st of that season.">Age</th>
  <th data-stat="team_id" align="left" class="tooltip sort_default_asc" tip="Team">Tm</th>
  <th data-stat="game_location" align="center" class="tooltip"></th>
  <th data-stat="opp_id" align="left" class="tooltip sort_default_asc" tip="Opponent">Opp</th>
  <th data-stat="game_result" align="center" class="tooltip"></th>
  <th data-stat="gs" align="right" class="tooltip" tip="Games Started">GS</th>
  <th data-stat="mp" align="right" class="tooltip" tip="Minutes Played">MP</th>
  <th data-stat="fg" align="right" class="tooltip" tip="Field Goals">FG</th>
  <th data-stat="fga" align="right" class="tooltip" tip="Field Goal Attempts">FGA</th>
  <th data-stat="fg_pct" align="right" class="tooltip" tip="Field Goal Percentage">FG%</th>
  <th data-stat="fg3" align="right" class="tooltip" tip="3-Point Field Goals">3P</th>
  <th data-stat="fg3a" align="right" class="tooltip" tip="3-Point Field Goal Attempts">3PA</th>
  <th data-stat="fg3_pct" align="right" class="tooltip" tip="3-Point Field Goal Percentage">3P%</th>
  <th data-stat="ft" align="right" class="tooltip" tip="Free Throws">FT</th>
  <th data-stat="fta" align="right" class="tooltip" tip="Free Throw Attempts">FTA</th>
  <th data-stat="ft_pct" align="right" class="tooltip" tip="Free Throw Percentage">FT%</th>
  <th data-stat="orb" align="right" class="tooltip" tip="Offensive Rebounds">ORB</th>
  <th data-stat="drb" align="right" class="tooltip" tip="Defensive Rebounds">DRB</th>
  <th data-stat="trb" align="right" class="tooltip" tip="Total Rebounds">TRB</th>
  <th data-stat="ast" align="right" class="tooltip" tip="Assists">AST</th>
  <th data-stat="stl" align="right" class="tooltip" tip="Steals">STL</th>
  <th data-stat="blk" align="right" class="tooltip" tip="Blocks">BLK</th>
  <th data-stat="tov" align="right" class="tooltip" tip="Turnovers">TOV</th>
  <th data-stat="pf" align="right" class="tooltip" tip="Personal Fouls">PF</th>
  <th data-stat="pts" align="right" class="tooltip" tip="Points">PTS</th>
  <th data-stat="game_score" align="right" class="tooltip" tip="Game Score">GmSc</th>
  <th data-stat="plus_minus" align="right" class="tooltip" tip="Plus/Minus">+/-</th>
...</table>

In []:

Deleting columns

In [38]:

del data['Rk']
del data['H/A']
del data['Tm']
#del data['Opp']
del data['GS']
del data['W/L']
data.head()

Out[38]:

	G	Date	Age	Opp	MP	FG	FGA	FG%	3P	3PA	3P%	FT	FTA	FT%	ORB	DRB	TRB	AST	STL	BLK	TOV	PF	PTS	GmSc	+/-
0	1	2012-11-01	24-033	SAS	40:39	9	18	0.500	1	2	0.500	4	5	0.800	2	12	14	5	2	0	4	1	23	19.7	2
1	2	2012-11-02	24-034	POR	42:41	7	14	0.500	1	3	0.333	8	12	0.667	2	15	17	7	0	2	6	1	23	20.2	13
2	3	2012-11-04	24-036	ATL	42:05	7	17	0.412	1	4	0.250	7	8	0.875	1	11	12	8	3	2	6	3	22	19.3	-8
3	4	2012-11-06	24-038	TOR	29:09	4	11	0.364	1	4	0.250	6	6	1.000	0	6	6	3	2	0	4	3	15	9.6	23
4	5	2012-11-08	24-040	CHI	38:23	11	19	0.579	0	2	0.000	2	2	1.000	1	3	4	1	3	3	6	1	24	16.1	-3

5 rows × 25 columns

Fields Goal Made/Fields Goal Attempted per minute¶

In [20]:

data[['MP','FG','FGA']].dtypes

Out[20]:

MP     object
FG      int64
FGA     int64
dtype: object

In [27]:

temp=data[['MP','FG','FGA']]

MP is object (string). It is not conviniet for us. Let us try to convert it into seconds...

In [22]:

import time
import datetime

In [26]:

def string_to_seconds(minutes):
    minutes=str(minutes)
    minutes=time.strptime(minutes,'%M:%S')
    return datetime.timedelta(minutes=minutes.tm_min, seconds=minutes.tm_sec).total_seconds()
    
print string_to_seconds('40:00')

2400.0

And now We can take every element of 'MP' column and... convert it to seconds

In [29]:

temp['MP']=temp['MP'].map(string_to_seconds)
temp.head()

-c:1: SettingWithCopyWarning: A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_index,col_indexer] = value instead

Out[29]:

	MP	FG	FGA
0	2439	9	18
1	2561	7	14
2	2525	7	17
3	1749	4	11
4	2303	11	19

5 rows × 3 columns

In [30]:

temp.dtypes

Out[30]:

MP     float64
FG       int64
FGA      int64
dtype: object

In [34]:

# Attempts per minute - create new column
temp['FGA/M']=temp['FGA']*60/temp['MP']
temp['FG/M']=temp['FG']*60/temp['MP']
temp.head()

-c:3: SettingWithCopyWarning: A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_index,col_indexer] = value instead

Out[34]:

	MP	FG	FGA	FG/M	FGA/M
0	2439	9	18	0.221402	0.442804
1	2561	7	14	0.163998	0.327997
2	2525	7	17	0.166337	0.403960
3	1749	4	11	0.137221	0.377358
4	2303	11	19	0.286583	0.495007

5 rows × 5 columns

In [35]:

temp.describe()

Out[35]:

	MP	FG	FGA	FG/M	FGA/M
count	81.000000	81.000000	81.000000	81.000000	81.000000
mean	2310.271605	9.024691	17.691358	0.234632	0.455967
std	346.315355	2.554287	5.001605	0.057639	0.094907
min	1426.000000	4.000000	8.000000	0.095390	0.267499
25%	2146.000000	7.000000	14.000000	0.202703	0.387812
50%	2346.000000	9.000000	17.000000	0.232288	0.447344
75%	2525.000000	10.000000	21.000000	0.267857	0.513619
max	2980.000000	16.000000	31.000000	0.411487	0.699029

8 rows × 5 columns

Let form new subset for groups¶

In [39]:

data.head()

Out[39]:

	G	Date	Age	Opp	MP	FG	FGA	FG%	3P	3PA	3P%	FT	FTA	FT%	ORB	DRB	TRB	AST	STL	BLK	TOV	PF	PTS	GmSc	+/-
0	1	2012-11-01	24-033	SAS	40:39	9	18	0.500	1	2	0.500	4	5	0.800	2	12	14	5	2	0	4	1	23	19.7	2
1	2	2012-11-02	24-034	POR	42:41	7	14	0.500	1	3	0.333	8	12	0.667	2	15	17	7	0	2	6	1	23	20.2	13
2	3	2012-11-04	24-036	ATL	42:05	7	17	0.412	1	4	0.250	7	8	0.875	1	11	12	8	3	2	6	3	22	19.3	-8
3	4	2012-11-06	24-038	TOR	29:09	4	11	0.364	1	4	0.250	6	6	1.000	0	6	6	3	2	0	4	3	15	9.6	23
4	5	2012-11-08	24-040	CHI	38:23	11	19	0.579	0	2	0.000	2	2	1.000	1	3	4	1	3	3	6	1	24	16.1	-3

5 rows × 25 columns

We deleted 'Opp' column earlier and have to rerun data=pd.read_csv('../data/KevinDuran2012-13GameLog.csv',names=columns) in the cell after 'Importing csv data' item of this paper...

In [40]:

group_by_opp=data.groupby('Opp')

In [42]:

group_by_opp.size()

Out[42]:

Opp
ATL    2
BOS    2
BRK    2
CHA    2
CHI    2
CLE    2
DAL    4
DEN    4
DET    2
GSW    4
HOU    3
IND    2
LAC    3
LAL    4
MEM    3
MIA    2
MIL    1
MIN    4
NOH    4
NYK    2
ORL    2
PHI    2
PHO    4
POR    4
SAC    3
SAS    4
TOR    2
UTA    4
WAS    2
dtype: int64

Now if you want figure out how mane shots he took against particular team

In [43]:

group_by_opp.sum()

Out[43]:

	G	FG	FGA	FG%	3P	3PA	3P%	FT	FTA	FT%	ORB	DRB	TRB	AST	STL	BLK	TOV	PF	PTS	GmSc	+/-
Opp
ATL	28	21	40	1.021	5	12	0.750	16	18	1.775	2	23	25	11	3	4	12	4	63	50.8	-6
BOS	76	15	36	0.825	3	11	0.429	19	19	2.000	0	14	14	6	2	1	6	6	52	35.5	8
BRK	50	20	33	1.210	4	7	1.400	15	16	1.750	0	10	10	11	3	1	6	7	59	49.1	1
CHA	77	12	20	1.250	3	7	0.833	10	11	1.857	1	11	12	11	0	5	4	2	37	37.8	60
CHI	61	17	38	0.895	1	6	0.250	8	8	2.000	3	17	20	7	3	5	10	1	43	31.4	19
CLE	54	17	37	0.944	4	9	0.929	20	25	1.640	2	17	19	5	1	1	5	3	58	42.4	16
DAL	183	42	89	1.954	13	24	2.456	45	46	3.833	2	34	36	11	7	6	17	7	142	106.8	26
DEN	206	33	75	1.818	6	25	0.950	44	49	3.484	2	28	30	20	10	5	18	10	116	90.0	19
DET	14	17	38	0.927	2	4	1.000	15	15	2.000	1	21	22	5	1	4	5	2	51	39.7	15
GSW	182	38	67	2.269	9	14	2.601	29	33	3.424	2	33	35	31	8	6	17	5	114	106.9	38
HOU	99	25	53	1.343	6	15	0.944	23	26	2.646	2	19	21	17	5	4	10	7	79	64.7	26
IND	97	22	45	0.994	3	8	0.867	14	17	1.657	2	15	17	5	0	0	6	1	61	40.1	22
LAC	113	29	63	1.400	10	20	1.541	34	39	2.705	3	19	22	16	8	0	14	4	102	78.9	18
LAL	161	45	92	1.960	9	23	1.597	40	46	3.429	3	25	28	18	4	4	8	12	139	106.4	56
MEM	124	33	65	1.589	2	7	1.250	25	28	2.678	3	22	25	14	4	4	11	5	93	71.8	4
MIA	80	23	45	1.024	3	9	0.700	24	26	1.818	1	14	15	7	1	3	9	10	73	49.8	-16
MIL	74	10	19	0.526	1	2	0.500	9	10	0.900	1	7	8	5	0	1	2	0	30	25.3	19
MIN	189	44	74	2.372	5	12	2.083	29	29	4.000	1	28	29	19	7	7	14	12	122	103.3	14
NOH	107	32	53	2.463	9	16	2.217	27	30	3.690	2	35	37	21	5	2	6	1	100	101.1	73
NYK	138	16	37	0.862	2	8	0.333	27	30	1.800	0	11	11	13	0	5	11	2	61	44.4	-2
ORL	136	15	29	1.033	3	8	0.867	18	23	1.568	2	15	17	5	6	2	11	4	51	38.9	-2
PHI	46	18	37	0.988	3	10	0.583	24	27	1.782	1	13	14	8	5	3	4	3	63	55.2	12
PHO	169	37	76	1.958	6	15	1.200	30	30	4.000	1	20	21	10	4	6	2	4	110	89.9	78
POR	190	36	63	2.295	5	16	1.222	19	24	3.542	4	32	36	21	3	6	16	5	96	80.6	52
SAC	148	27	42	1.922	6	12	1.517	24	26	2.750	1	22	23	16	8	3	9	6	84	81.8	41
SAS	164	32	64	2.000	4	8	2.167	25	27	3.550	3	32	35	20	5	8	19	6	93	75.1	7
TOR	37	10	22	0.909	3	9	0.650	14	15	1.889	1	12	13	10	3	1	5	4	37	33.6	51
UTA	212	32	52	2.441	5	9	1.750	33	37	3.596	0	32	32	17	7	7	17	9	102	89.6	34
WAS	106	13	29	0.874	4	8	1.333	19	20	1.900	0	13	13	14	3	1	6	1	49	44.5	32

29 rows × 21 columns

In [44]:

data[data.Opp=='ATL']

Out[44]:

	G	Date	Age	Opp	MP	FG	FGA	FG%	3P	3PA	3P%	FT	FTA	FT%	ORB	DRB	TRB	AST	STL	BLK	TOV	PF	PTS	GmSc	+/-
2	3	2012-11-04	24-036	ATL	42:05	7	17	0.412	1	4	0.25	7	8	0.875	1	11	12	8	3	2	6	3	22	19.3	-8
24	25	2012-12-19	24-081	ATL	40:52	14	23	0.609	4	8	0.50	9	10	0.900	1	12	13	3	0	2	6	1	41	31.5	2

2 rows × 25 columns

In [62]:

field_goal_per_team=group_by_opp.sum()[['FGA','FG']]
field_goal_per_team

Out[62]:

	FGA	FG
Opp
ATL	40	21
BOS	36	15
BRK	33	20
CHA	20	12
CHI	38	17
CLE	37	17
DAL	89	42
DEN	75	33
DET	38	17
GSW	67	38
HOU	53	25
IND	45	22
LAC	63	29
LAL	92	45
MEM	65	33
MIA	45	23
MIL	19	10
MIN	74	44
NOH	53	32
NYK	37	16
ORL	29	15
PHI	37	18
PHO	76	37
POR	63	36
SAC	42	27
SAS	64	32
TOR	22	10
UTA	52	32
WAS	29	13

29 rows × 2 columns

In [63]:

stacked=vincent.StackedBar(field_goal_per_team)

In [64]:

stacked.legend(title='Field Goals')
stacked.scales['x'].padding=0.1
display(stacked)

<vincent.charts.Bar at 0xb66af98>

In [66]:

stacked.display()

<IPython.core.display.Javascript at 0xbea6588>

In []:

Посты чуть ниже также могут вас заинтересовать

iPython R Rapid Miner

Поиск по блогу

Страницы

пятница, 14 февраля 2014 г.

Pandas Data Frames and Kevin Durant 2012-13 Game Log with Mahdi Yusuf

Introductijn Pandas DataFrames¶

Importing csv data¶

Fields Goal Made/Fields Goal Attempted per minute¶

Let form new subset for groups¶

Комментариев нет:

Отправить комментарий

	FGA	FG
Opp
ATL	40	21
BOS	36	15
BRK	33	20
CHA	20	12
CHI	38	17
CLE	37	17
DAL	89	42
DEN	75	33
DET	38	17
GSW	67	38
HOU	53	25
IND	45	22
LAC	63	29
LAL	92	45
MEM	65	33
MIA	45	23
MIL	19	10
MIN	74	44
NOH	53	32
NYK	37	16
ORL	29	15
PHI	37	18
PHO	76	37
POR	63	36
SAC	42	27
SAS	64	32
TOR	22	10
UTA	52	32
WAS	29	13

	FGA	FG
Opp
ATL	40	21
BOS	36	15
BRK	33	20
CHA	20	12
CHI	38	17
CLE	37	17
DAL	89	42
DEN	75	33
DET	38	17
GSW	67	38
HOU	53	25
IND	45	22
LAC	63	29
LAL	92	45
MEM	65	33
MIA	45	23
MIL	19	10
MIN	74	44
NOH	53	32
NYK	37	16
ORL	29	15
PHI	37	18
PHO	76	37
POR	63	36
SAC	42	27
SAS	64	32
TOR	22	10
UTA	52	32
WAS	29	13

Поиск по блогу

Страницы

пятница, 14 февраля 2014 г.

Pandas Data Frames and Kevin Durant 2012-13 Game Log with Mahdi Yusuf

Introductijn Pandas DataFrames¶

Importing csv data¶

Fields Goal Made/Fields Goal Attempted per minute¶

Let form new subset for groups¶

Комментариев нет:

Отправить комментарий

пятница, 14 февраля 2014 г.

	FGA	FG
Opp
ATL	40	21
BOS	36	15
BRK	33	20
CHA	20	12
CHI	38	17
CLE	37	17
DAL	89	42
DEN	75	33
DET	38	17
GSW	67	38
HOU	53	25
IND	45	22
LAC	63	29
LAL	92	45
MEM	65	33
MIA	45	23
MIL	19	10
MIN	74	44
NOH	53	32
NYK	37	16
ORL	29	15
PHI	37	18
PHO	76	37
POR	63	36
SAC	42	27
SAS	64	32
TOR	22	10
UTA	52	32
WAS	29	13