def display_pair_plot(fileName=None):
#consider two versions of the data :
#dict the whole dataset with keys = columns names, values = list columns values
#dict_per_house with only columns of interest (with marks), keys = column name, val = dict with key = house, values = columns values for this house
dict = getDataSet(fileName)
dict_per_house = distribute_per_house(dict)
#n_rows is the number of key ie the number of courses
n_rows = len(dict_per_house)
#we define a big structure for the plot
f, ax = plt.subplots(n_rows, n_rows, figsize=(50, 75))
list_keys = list(dict_per_house.keys())
#we scan all keys
for i in range(n_rows):
for j in range(n_rows):
#x_name is the name of the course x
#y_name is the name of the course y
x_name = list_keys[i]
y_name = list_keys[j]
#if different, we plot a scatter plot
if (x_name != y_name):
plot_scatter(ax[i, j], x_name, y_name, dict_per_house)
#if identical, we plot an histogram
if (x_name == y_name):
plot_hist(ax[i, j], x_name, dict, dict_per_house, 50)
plt.tight_layout()
#we save the final plot in a .png file
f.savefig("pair_plot.png")
def test_equalize_hist(filename, dist): im = Image.open(filename) result = equalize_hist(im) out_path = os.path.join(dist, 'equalized.png') result.save(out_path) print 'The equalized image has been saved to %s.' % out_path # new histogram plot_hist(result, os.path.join(dist, 'equalized_hist.png'))
def hist_tot(feature_dico, feature_list, b, house_colors):
l = 0
m = 0
fig = plt.figure(figsize=(20, 10), dpi=100)
grid = gridspec.GridSpec(4, 4)
for k in range(len(feature_list)):
xy_dico = freq_per_house(feature_dico[feature_list[k]], b)
if k < 4:
l = k
else:
m = k // 4
l = k - 4 * m
plt.subplot(grid[m, l])
plot_hist(feature_dico, feature_list[k], b, house_colors)
plt.title(feature_list[k])
plt.legend(ncol=2, fontsize='x-small')
plt.tight_layout()
plt.show(block=True)
def pair_plot(feature_dico, feature_list, house_colors):
n = len(feature_list)
fig = plt.figure(figsize=(15, 15), dpi=100)
grid = gridspec.GridSpec(n, n)
for k in range(n):
for l in range(n):
plt.subplot(grid[k, l])
if k == l:
plot_hist(feature_dico, feature_list[k], 20, house_colors)
if k == 0:
plt.legend(ncol=2, fontsize='x-small')
else:
plot = scatter_plot(feature_dico, feature_list[k],
feature_list[l], house_colors)
if k == 0:
plt.title(feature_list[l])
if l == 0:
plt.ylabel(feature_list[k])
plt.tight_layout()
plt.show(block=True)
