Generate a heatmap in MatPlotLib

Question

I've been attempting to generate a heatmap by following the top answer over here: Generate a heatmap in MatPlotLib using a scatter data set

With some help from you kind folk here I've shaped my data so I can mine with relative ease (though not very pretty code) the x and y coordinates.

The coordinates are put nicely into arrays but when I plug those arrays into the heatmap code I get the following output:

The code I'm using is as follows:

import csv
import string, numpy as np
import numpy as np
import numpy.random
import matplotlib.pyplot as plt
import pickle
import operator as oper
import re
from math import sqrt
import re # used here to remove non-numeric values



file = open('P2E2D_Long_Format.csv', 'rb')

reader1 = csv.reader(file, delimiter = '\t')
non_decimal = re.compile(r'[^\d.]+')

x = []
y = []

count = 0 

for row in reader1:
    #print row
    if 'SCL' in row[0] and count <= 5000:
        #print row[0]
        if len(row[2]) and len(row[3]) > 0: #ensures that x and y CoOr are not empty.
            if '   .' not in row[2] and 'SAMPLES' not in row[3]:
                #populates list with values in row 2 & 3            
                xCoOr = row[2]
                yCoOr = row[3]

                #this code removes the blank spaces before numbers
                x1 = non_decimal.sub('', xCoOr)
                y1 = non_decimal.sub('', yCoOr)
                print x1, 'XCoOr'
                print y1, 'YCoOr'


                #changes values from file, that are seen as string, to a float            
                x1 = float(x1)
                y1 = float(y1)
                #print x1            
                x = x + [x1]
                y = y + [y1]
                count = count + 1
                print count



myarrayx = np.asarray(x)
myarrayy = np.asarray(y)

myarrayx = myarrayx - 508
myarrayy = myarrayy - 384

myarrayx = myarrayx / 100
myarrayy = myarrayy / 100


heatmap, xedges, yedges = np.histogram2d(myarrayx, myarrayy, bins=50)
extent = [xedges[0], xedges[-1], yedges[0], yedges[-1]]
plt.clf()
plt.imshow(heatmap, extent=extent)
plt.show()

file.close()   # <---IMPORTANT
#Long_Format.close()
#textFile.close()

row[0] is the Trial ID and row2 is the x CoOr and row[3] the y CoOr.

When looking at the outputted figure above it appears as though the y-axis is completely absent.

Does anyone have any ideas?

Here is some of the data I'm working with:

"Trial ID   Frame ID                    X-CoOr  Y-CoOr  Time"


"SCL    5413917 PreBeep1_1st_Sketchpad  653.8   542.1   4844"

"SCL    5413917 PreBeep1_1st_Sketchpad  654.7   542.2   4847"

"SCL    5413919 order of frames:        655.5   541.9   4849"

"SCL    5413919 order of frames:        656.1   541.2   4851"

"SCL    5413921 crosshair               655.8   540.8   4851"

"SCL    5413921 crosshair               655.7   540.6   4847"

"SCL    5413923 sketchpad               655.7   540.6   4843"

"SCL    5413923 sketchpad               655.5   540.6   4838"

"SCL    5413925 sketchpad               655.3   540.7   4838"

"SCL    5413925 sketchpad               655.1   540     4833"

"SCL    5413927 sketchpad               655.3   538.9   4829"

"SCL    5413927 sketchpad               655.4   538.1   4825"

"SCL    5413929 buffer1                 655.6   537.8   4824"

"SCL    5413929 buffer1                 655.5   537.5   4824"

"SCL    5413931 Diode1                  655.2   537.3   4824"

"SCL    5413931 Diode1                  654.9   537.6   4831"

"SCL    5413931 Diode1                  654.9   538.1   4836"

"SCL    5413931 Diode1                  654.8   538.6   4841"

"SCL    5413931 Diode1                  654.8   539 4841"

"SCL    5413931 Diode1                  655.6   539.1   4841"

EDIT: I have manually manipulated the min and max values on the x and y axis in the figure window as follows:

As can be seen the x-axis max value is huge. I'm not sure why this is. I suspect this value must appear in the xCoOr column in my data.

However, even when I adjust max and min values the heatmap is blank.

Thanks for your time on this.

Answer 1

My colleague solved the issue.

The problem was with the aspect ratio.

I first filtered out any x or y values above 2000 (as there was some enormous artefacts in the data).

My colleague then wrote the following code:

plt.imshow(heatmap, extent=extent, aspect='auto')

The key being he added

aspect='auto'

This stretched the x-axis to match the y-axis and thus made everything more visible.

Corrected output:

Thanks to everyone who considered the problem.