CODEWITHSUNDEEP
pythonpython-3.xmatplotlibboxplot
Hernan
Hernan

Reputation: 1148

First boxplot missing or not visible

Why is the first boxplot missing? There's supoussed to be 24 boxplots, but only 23 are being display as you can see in the image. Who's size do i have to change to make it visible? I've tried changing the size of the figure but its the same.

Enero Boxplot is out of the square figure

Not sure if it helps, but this is the code:

def obtenerBoxplotsAnualesIntercalados(self, directorioEntrada, directorioSalida):
    meses = ["Enero","Febrero","Marzo","Abril","Mayo","Junio", "Julio", "Agosto","Septie.","Octubre","Noviem.","Diciem."]
    ciudades = ["CO","CR"]      
    anios = ["2011", "2012", "2013"]
    for anio in anios:
        fig = plt.figure()
        fig.set_size_inches(14.3, 9)
        ax = plt.axes()
        plt.hold(True)
        bpArray = []
        i=0
        ticks = []
        for mes in range(len(meses)):
            archivoCO = open(directorioEntrada+"/"+"CO"+"-"+self.mesStr(mes+1)+"-"+anio, encoding = "ISO-8859-1")
            archivoCR = open(directorioEntrada+"/"+"CR"+"-"+self.mesStr(mes+1)+"-"+anio, encoding = "ISO-8859-1")
            datosCOmes = self.obtenerDatosmensuales(archivoCO)
            datosCRmes = self.obtenerDatosmensuales(archivoCR)
            data = [    [int(float(datosCOmes[2])), int(float(datosCOmes[0])), int(float(datosCOmes[1]))],
                        [int(float(datosCRmes[2])), int(float(datosCRmes[0])), int(float(datosCRmes[1]))] ]
            bpArray.append(plt.boxplot(data, positions=[i,i+1], widths=0.5))
            ticks.append(i+0.5)
            i=i+2
        hB, = plt.plot([1,1],'b-')
        hR, = plt.plot([1,1],'r-')
        plt.legend((hB, hR),('Caleta', 'Comodoro'))
        hB.set_visible(False)
        hR.set_visible(False)
        ax.set_xticklabels(meses)
        ax.set_xticks(ticks)
        self.setBoxColors(bpArray)
        plt.title('Variación de temperatura mensual Caleta Olivia-Comodoro Rivadavia. Año '+anio)
        plt.savefig(directorioSalida+"/asdasd"+str(anio)+".ps", orientation='landscape', papertype='A4' )

Upvotes: 0

Views: 568

Answers (1)

areuexperienced
areuexperienced

Reputation: 2071

Your boxplot is there but it's hidden. This reproduces your problem:

import matplotlib
import numpy as np

data = np.random.normal(10,2,100*24).reshape(24,-1) # let's get 12 pairs of arrays to plot
meses = ["Enero","Febrero","Marzo","Abril","Mayo","Junio", "Julio", "Agosto","Septie.","Octubre","Noviem.","Diciem."]

ax = plt.axes()
plt.hold(True)

i=0

ticks = []

for mes in range(0,len(meses)):
    plt.boxplot(data, positions=[i,i+1], widths=0.5)
    ticks.append(i+0.5)
    i+=2

ax.set_xticklabels(meses)
ax.set_xticks(ticks)    

plt.show()

Notice that you are defining your positions as ranging from 0 to 12, but you append ticks as range(0,12) + 0.5. Thus, when you later do set_xticks(ticks), your x axis will begin from 0.5 but your 1st boxplot is plotted at position 0.

wrong

I've adapted your code slightly to produce the result you want:

ax = plt.axes()
plt.hold(True)

i=1 # we start plotting from position 1 now

ticks = []

for mes in range(0,len(meses)):
    plt.boxplot(data, positions=[i,i+1], widths=0.5)
    ticks.append(i+0.5)
    i+=2

ax.set_xticklabels(meses)
ax.set_xlim(0,ticks[-1]+1) # need to shift the right end of the x limit by 1
ax.set_xticks(ticks)    


plt.show()

fixed

Upvotes: 1

Related Questions