CODEWITHSUNDEEP
pythonpandas
Greg Milani
Greg Milani

Reputation: 33

Graphing Certain DataPoints from Pandas

I'm trying to build a program that deals with baseball statistics. I Ask the user to input a team and then the code runs through a panda I have created searching for "teamID" that matches the user input.

I've tried grouping by "teamID" but and indexing before the for loop.

def AttendancePlot(teams,team_pick):

    fig, ax = plt.subplots()
    group_by_teamID = teams.groupby(by=['teamID'])
    print group_by_teamID

    for i in group_by_teamID.index:
        if i == team_pick:
            ax.scatter(teams['yearID'][i], teams['attendance'][i], color="#4DDB94", s=200)
            ax.annotate(i, (teams['yearID'][i], teams['attendance'][i]),
               bbox=dict(boxstyle="round", color="#4DDB94"),
               xytext=(-30, 30), textcoords='offset points',
               arrowprops=dict(arrowstyle="->", connectionstyle="angle,angleA=0,angleB=90,rad=10"))

How I'm creating the Panda

teams = pd.read_csv('Teams.csv')
salaries = pd.read_csv('Salaries.csv')
names = pd.read_csv('Names.csv')

teams = teams[teams['yearID'] >= 1985]
teams = teams[['yearID', 'teamID', 'Rank', 'R', 'RA', 'G', 'W', 'H', 'BB',    'HBP', 'AB', 'SF', 'HR', '2B', '3B', 'attendance']]
teams = teams.set_index(['yearID', 'teamID'])

salaries_by_yearID_teamID = salaries.groupby(['yearID', 'teamID'])  ['salary'].sum()
teams = teams.join(salaries_by_yearID_teamID)

print teams.head(15)

Outputted Panda 

          Rank    R   RA    G     ...       2B  3B  attendance      salary
yearID teamID                          ...                                     
1985   ATL        5  632  781  162     ...      213  28   1350137.0   14807000.0
       BAL        4  818  764  161     ...      234  22   2132387.0  11560712.0
       BOS        5  800  720  163     ...      292  31   1786633.0  10897560.0
       CAL        2  732  703  162     ...      215  31   2567427.0  14427894.0

I would like a scatter plot showing yearly attendance of a certain inputted team. I am getting a blank graph with no errors.

Upvotes: 3

Views: 62

Answers (1)

Valentino
Valentino

Reputation: 7361

No need to use groupby() here, groupby() is typically used when you want to apply some math on a selection of rows. What you need is a proper selection of the data.

This function will plot year (x axis) vs attendance (y axis) of the given team team_pick, assuming the dataframe structure you described (dataframe is teams):

def AttendancePlot(teams, team_pick):
    teamdata = teams.loc[teams.index.get_level_values('teamID') == team_pick]
    plt.scatter(teamdata.index.levels[0], teamdata['attendance'])
    plt.show()

I leave annotation to you.

The key is this line: teamdata = teams.loc[teams.index.get_level_values('teamID') == team_pick].
teams.index.get_level_values('teamID') == team_pick performs a selection on the multiline index, allowing you to select all rows where the team is team_pick.
teamdata is hence a dataframe containing all the rows for the given team.

This is called pandas indexing. See also the pandas advanced indexing.

Upvotes: 2

Related Questions