💄 Fixes minor mistakes

analyze.py

@@ -119,10 +119,14 @@ def plot_violin(data, labels, Fs, ps, title):
             axs[j, k].set_ylabel(grade_name_labels[index], fontweight="bold")
 
             # q1-q3 lines
-            for vec in data[index]:
+            for ind, vec in enumerate(data[index]):
-                inds = np.arange(1, len(data[index]) + 1)
                 quartile1, median, quartile3 = np.percentile(vec, [25, 50, 75])
-                axs[j, k].vlines(inds, quartile1, quartile3, color="gray", linewidths=3)
+                if quartile1 == quartile3:
+                    if quartile1 >= 0.1:
+                        quartile1 -= 0.1
+                    if quartile3 <= max(vec) - 0.1:
+                        quartile3 += 0.1
+                axs[j, k].vlines(ind + 1, quartile1, quartile3, color="gray", linewidths=3)
 
             axs[j, k].set_xticks(np.arange(1, len(labels) + 1), labels=labels)
             axs[j, k].set_yticks(np.arange(1, 5.01, step))
@@ -134,9 +138,9 @@ def plot_violin(data, labels, Fs, ps, title):
                 part.set_facecolor(colors[i % len(colors)])
                 part.set_edgecolor(edge_colors[i % len(edge_colors)])
 
-            F = round(Fs[index], 2)
+            F = Fs[index]
-            p = round(ps[index], 4)
+            p = ps[index]
-            axs[j, k].text(0.01, 0.99, f"F-stat: {F:.2f}\np-val: {p:.4f}", ha="left", va="top",
+            axs[j, k].text(0.01, 0.99, f"F-stat: {F:.4f}\np-val: {p:.4f}", ha="left", va="top",
                            transform=axs[j, k].transAxes,
                            fontweight="bold")
             axs[j, k].text(0.99, 0.99,
@@ -147,12 +151,12 @@ def plot_violin(data, labels, Fs, ps, title):
 
             medians = list([np.median(a) for a in data[index]])
             means = list([a.mean() for a in data[index]])
-            for l in range(len(medians)):
+            for l in range(len(data[index])):
-                median = round(medians[l], 2)
+                median = medians[l]
-                mean = round(means[l], 2)
+                mean = means[l]
                 # left - mean, right - median
-                axs[j, k].text(l + 1.13, median - 0.05, f"{median}", color="green")
+                axs[j, k].text(l + 1.13, median - 0.05, f"{median:.2f}", color="green")
-                axs[j, k].text(l + 0.90 - len(labels) * 0.065, mean - 0.05, f"{mean}", color="red")
+                axs[j, k].text(l + 0.90 - len(labels) * 0.065, mean - 0.05, f"{mean:.2f}", color="red")
 
     fig.tight_layout()
     if save != "":

analyze_absence.py

@@ -11,6 +11,9 @@ args = parser.parse_args()
 graph = args.graph
 save = args.save
 
+colors = ["lightblue", "lightgreen", "lightcoral"]
+edge_colors = ["blue", "green", "red"]
+
 dataset = np.load("clean.npy")
 print(f"dataset shape: {dataset.shape}; analyzing column 11 (absence)")
 print("\tinteger value")
@@ -55,29 +58,71 @@ for j in range(2):
         step = 1 if index > 0 else 0.5
 
         if index == 0:
-            axs[j, k].scatter(dataset[:, 11], dataset[:, 2])
+            x = dataset[:, 11]  # absence
+            y = dataset[:, 2]  # gpa
+            axs[j, k].scatter(x, y)
             axs[j, k].set_xlabel("Počet vymeškaných hodín")
             axs[j, k].set_ylabel(grade_name_labels[index])
+
+            # trendline
+            z = np.polyfit(x, y, 1)
+            p = np.poly1d(z)
+
+            axs[j, k].plot(x, p(x), color="gray")
         else:
             current = list([data[index][0][data[index][1] == i + 1] for i in range(5)])  # i wanna kms
-            axs[j, k].violinplot(list(filter(lambda x: len(x), current)), showmeans=True)
+            # data[index][0] = absence
-            axs[j, k].set_xticks(np.arange(1, 6, 1), labels=["1", "2", "3", "4", "5"])
+            # data[index][1] = grade
+            # data[index][0][where this specific grade] -> absences for that sepcific grade
+            # iterate 1 through 5 and plug into ^^
+
+            parts = axs[j, k].violinplot(list([x if len(x) else [0] for x in current]), showmeans=True,
+                                         showmedians=True)
+            axs[j, k].set_xticks(np.arange(1, 6), labels=["1", "2", "3", "4", "5"])
             axs[j, k].set_xlabel(grade_name_labels[index])
             axs[j, k].set_ylabel("Počet vymeškaných hodín")
 
+            # q1-q3 lines
+            for ind, vec in enumerate(current):
+                if len(vec) == 0:
+                    continue
+                quartile1, median, quartile3 = np.percentile(vec, [25, 50, 75])
+                print(quartile1, median, quartile3)
+                axs[j, k].vlines(ind + 1, quartile1, quartile3, color="gray", linewidths=3)
+
+            parts["cmeans"].set_color("red")
+            parts["cmedians"].set_color("green")
+
+            for i, part in enumerate(parts["bodies"]):
+                part.set_facecolor(colors[i % len(colors)])
+                part.set_edgecolor(edge_colors[i % len(edge_colors)])
+
         axs[j, k].set_title(grade_names[index])
 
-        tau = round(taus[index], 2)
+        tau = taus[index]
-        p = round(ps[index], 4)
+        p = ps[index]
-        axs[j, k].text(0.01, 0.99, f"Tau τ: {tau:.2f}\np-val: {p:.4f}", ha="left", va="top", transform=axs[j, k].transAxes,
+        axs[j, k].text(0.01, 0.99, f"Tau τ: {tau:.4f}\np-val: {p:.4f}", ha="left", va="top",
+                       transform=axs[j, k].transAxes,
                        fontweight="bold")
 
         if index:
-            by_grade = [data[index][0][data[index][1] == i + 1] for i in range(5)]
+            axs[j, k].text(0.99, 0.99,
-            means = list([a.mean() for a in filter(lambda b: len(b), by_grade)])
+                           f"Na ľavo - priemer (červená)\nNa pravo - medián (zelená)\nSivá - medzi kvartilom 1 a 3",
-            for l in range(len(means)):
+                           ha="right",
-                mean = round(means[l], 2)
+                           va="top",
-                axs[j, k].text(l + 1.02, mean + 5, f"{mean}")
+                           transform=axs[j, k].transAxes)
+
+            current = list([data[index][0][data[index][1] == i + 1] for i in range(5)])  # i wanna kms
+            medians = list([np.median(a) if len(a) else -1 for a in current])
+            means = list([a.mean() if len(a) else -1 for a in current])
+            for l in range(len(current)):
+                median = medians[l]
+                mean = means[l]
+                # left - mean, right - median
+                if median >= 0:
+                    axs[j, k].text(l + 1.14, median - 5, f"{median:.2f}", color="green", ha="left")
+                if mean >= 0:
+                    axs[j, k].text(l + 0.85, mean - 5, f"{mean:.2f}", color="red", ha="right")
 
 fig.tight_layout()
 if save != "":