Summary plots py

summary_plots.py

+import matplotlib.pyplot as plt
+import json
+import numpy as np
+import pandas as pd
+from datetime import datetime
+from sklearn.metrics import mean_squared_error
+import statistics
+from scipy.stats import norm 
+import scipy.stats as stats
+
+def read_moodle_json(file_path, key_list):
+    try:
+        with open(file_path, 'r') as file:
+            data = json.load(file)[0]
+            responses = { }
+            for item in data:
+                neptun = item["idnumber"]
+                responses[neptun] = { }
+                for q in key_list:
+                    responses[neptun][q] = item[q]
+            return responses
+    except FileNotFoundError:
+        print(f"Error: The file {file_path} was not found.")
+    except json.JSONDecodeError:
+        print(f"Error: The file {file_path} contains invalid JSON.")
+
+ZH1_A_responses = read_moodle_json("BMEVIIIAA03_HU-1.NZHAcsoport-responses.json", 
+                            ["response12", "response13", "response14", "response15", "response16"])
+ZH1_A_grades= read_moodle_json("BMEVIIIAA03_HU-1.NZHAcsoport-grades.json", 
+                            ["grade4000", "q12300", "q13500", "q14400", "q15300", "q16200"])
+
+teachers = ["Zoltán István Karsa", "..."]
+
+graders = { }
+
+def rename_key(old_key, extra_key, except_keys = ["name", "username", "idnumber", "emailaddress"]):
+    if old_key in except_keys:
+        return old_key
+    return extra_key + old_key
+
+def rename_dict(dict, extra_key):
+    new = { }
+    for i in dict.keys():
+        key = rename_key(i, extra_key)
+        new[key] = dict[i]
+    return new
+
+def read_json_to_pandas(file_path, assigment = ""):
+    try:
+        with open(file_path, 'r') as file:
+            data = json.load(file)[0]
+            filtered = [d for d in data if d['source'] == "mod/quiz" and d["name"] not in teachers and d["revisedgrade"] != ""]
+            sorted_data = sorted(
+                filtered, 
+                key=lambda x: datetime.strptime(x["dateandtime"], "%A, %d %B %Y, %I:%M %p"), reverse=False
+            )
+            filt_duplicate = {  }
+            for i in sorted_data:
+                if i["username"] not in filt_duplicate:
+                    i["dateandtime"] = datetime.strptime(i["dateandtime"], "%A, %d %B %Y, %I:%M %p")
+                    i["originalgrade"] = float(i["revisedgrade"])
+                    i["revisedgrade"] = float(i["revisedgrade"])
+                    i["grader"] = "student"
+                    filt_duplicate[i["username"]] = rename_dict(i, assigment)
+                else:
+                    filt_duplicate[i["username"]][rename_key("revisedgrade", assigment)] = float(i["revisedgrade"])
+                    filt_duplicate[i["username"]][rename_key("overridden", assigment)] = "Yes"
+                    filt_duplicate[i["username"]][rename_key("grader", assigment)] = i["grader"]
+                    if i["grader"] not in graders:
+                        graders[i["grader"]] = 1
+                    else:
+                        graders[i["grader"]] = graders[i["grader"]] + 1
+
+            dataframe = pd.DataFrame(filt_duplicate.values())
+            dataframe = dataframe.set_index("username")
+            dataframe[rename_key("diff", assigment)] = abs(dataframe[rename_key("revisedgrade", assigment)] - dataframe[rename_key("originalgrade", assigment)])
+            return dataframe
+    except FileNotFoundError:
+        print(f"Error: The file {file_path} was not found.")
+    except json.JSONDecodeError:
+        print(f"Error: The file {file_path} contains invalid JSON.")
+
+df = pd.concat([read_json_to_pandas("2024ZH1A.json", "ZH1-"), read_json_to_pandas("2024ZH1B.json", "ZH1-")])
+df2 = pd.concat([read_json_to_pandas("2024ZH2A.json", "ZH2-"), read_json_to_pandas("2024ZH2B.json", "ZH2-")])
+df3 = pd.concat([read_json_to_pandas("2024PZH1.json", "PZH1-")])
+df4 = pd.concat([read_json_to_pandas("2024PZH2.json", "PZH2-")])
+df = pd.concat([df, df2, df3, df4], axis=1)
+#df.to_excel("output.xlsx")
+
+ZH1_over_notnull = df[(df[rename_key("overridden", "ZH1-")] == "Yes") & df[[rename_key("originalgrade", "ZH1-"), rename_key("revisedgrade", "ZH1-")]].notnull().all(1)]
+ZH2_over_notnull = df[(df[rename_key("overridden", "ZH2-")] == "Yes") & df[[rename_key("originalgrade", "ZH2-"), rename_key("revisedgrade", "ZH2-")]].notnull().all(1)] 
+ZH1_notnull = df[df[[rename_key("originalgrade", "ZH1-"), rename_key("revisedgrade", "ZH1-")]].notnull().all(1)]
+ZH2_notnull = df[df[[rename_key("originalgrade", "ZH2-"), rename_key("revisedgrade", "ZH2-")]].notnull().all(1)] 
+
+ZH_notnull = df[df[[rename_key("originalgrade", "ZH1-"), rename_key("revisedgrade", "ZH1-"), rename_key("originalgrade", "ZH2-"), rename_key("revisedgrade", "ZH2-")]].notnull().all(1)]
+
+
+Z, xedges, yedges = np.histogram2d(ZH_notnull[rename_key("originalgrade", "ZH1-")], ZH_notnull[rename_key("originalgrade", "ZH2-")], bins=40, range=[[0, 40],[0, 40]])
+Z2, xedges, yedges = np.histogram2d(ZH_notnull[rename_key("revisedgrade", "ZH1-")], ZH_notnull[rename_key("revisedgrade", "ZH2-")], bins=40, range=[[0, 40],[0, 40]])
+
+im = plt.pcolormesh(xedges, yedges, Z2-Z, shading='flat', 
+                    #cmap = plt.colormaps['Greys']
+                    )
+plt.colorbar(im)
+plt.axline((0,40), slope=-1, c="black", linestyle='dashed', label="Min. Req")
+plt.xlabel("ZH1 grade")
+plt.ylabel("ZH2 grade")
+#plt.legend()
+plt.show()
+
+mu, std = norm.fit(ZH1_notnull[rename_key("originalgrade", "ZH1-")])
+mean_revised_zh1 = statistics.mean(ZH1_notnull[rename_key("revisedgrade", "ZH1-")])
+sd_revised_zh1 = statistics.stdev(ZH1_notnull[rename_key("revisedgrade", "ZH1-")])
+
+x_axis = np.arange(0, 40, 0.01) 
+
+plt.hist([df[rename_key("originalgrade", "ZH1-")], df[rename_key("revisedgrade", "ZH1-")], 
+          #df[rename_key("originalgrade", "ZH2-")], df[rename_key("revisedgrade", "ZH2-")]
+            ], 
+          label=["Original grade ZH1", "Revised grade ZH1", 
+            #     "Original grade ZH2", "Revised grade ZH2"
+            ], 
+          color=["red", "blue", 
+                 #"blue", "cornflowerblue"
+                 ], bins=20, )
+plt.plot(x_axis, norm.pdf(x_axis, mu, std) * len(df[rename_key("originalgrade", "ZH1-")]), label="Original N", color="red")
+mu, std = norm.fit(ZH1_notnull[rename_key("revisedgrade", "ZH1-")])
+plt.plot(x_axis, norm.pdf(x_axis, mu, std) * len(df[rename_key("revisedgrade", "ZH1-")]), label="Revised N", color="blue")
+plt.xlabel('Point')
+plt.ylabel('Frequency')
+plt.legend()
+plt.show()
+
+counts, bins, bars = plt.hist([ZH1_over_notnull[rename_key("originalgrade", "ZH1-")], ZH1_over_notnull[rename_key("revisedgrade", "ZH1-")], 
+          ZH2_over_notnull[rename_key("originalgrade", "ZH2-")], ZH2_over_notnull[rename_key("revisedgrade", "ZH2-")]], 
+          label=["Original grade ZH1", "Revised grade ZH1", "Original grade ZH2", "Revised grade ZH2"], 
+          color=["red", "lightcoral", "blue", "cornflowerblue"], bins=10, 
+          #density=True
+          )
+
+count = len(ZH1_over_notnull[rename_key("originalgrade", "ZH1-")])
+
+mu, std = norm.fit(ZH1_over_notnull[rename_key("originalgrade", "ZH1-")])
+plt.plot(x_axis, norm.pdf(x_axis, mu, std)*count, label="Original ZH1 N", color="red")
+
+density = [sum(counts[0][0:3])]
+density.extend(counts[0][3:])
+density = np.array(density)
+print(density)
+limits = [16, 20, 24, 28, 32, 36, 40]
+expected = norm.cdf(limits, mu, std) * float(count)
+propability = [expected[0]]
+propability.extend(expected[1:] - expected[0:-1])
+propability.append(count - expected[-1])
+print(propability)
+
+test_stat, p_value = stats.chisquare(density, propability)
+# chi square test statistic and p value 
+print('chi_square_test_statistic is : ' +
+      str(test_stat)) 
+print('p_value : ' + str(p_value)) 
+print(stats.chi2.ppf(1-0.05, df=14)) 
+
+mu, std = norm.fit(ZH1_over_notnull[rename_key("revisedgrade", "ZH1-")])
+plt.plot(x_axis, norm.pdf(x_axis, mu, std)*count, label="Revised ZH1 N", color="lightcoral")
+mu, std = norm.fit(ZH2_over_notnull[rename_key("originalgrade", "ZH2-")])
+plt.plot(x_axis, norm.pdf(x_axis, mu, std), label="Original ZH2 N", color="blue")
+mu, std = norm.fit(ZH2_over_notnull[rename_key("revisedgrade", "ZH2-")])
+plt.plot(x_axis, norm.pdf(x_axis, mu, std), label="Revised ZH2 N", color="cornflowerblue")
+plt.xlabel('Point')
+plt.ylabel('Density')
+plt.legend()
+plt.show()
+
+
+counts, bins, bars = plt.hist([ZH1_over_notnull[rename_key("originalgrade", "ZH1-")], ZH1_over_notnull[rename_key("revisedgrade", "ZH1-")], 
+          ZH2_over_notnull[rename_key("originalgrade", "ZH2-")], ZH2_over_notnull[rename_key("revisedgrade", "ZH2-")]], 
+          label=["Original grade ZH1", "Revised grade ZH1", "Original grade ZH2", "Revised grade ZH2"], 
+          color=["red", "lightcoral", "blue", "cornflowerblue"], bins=20, 
+          #density=True
+          )
+
+count = len(ZH1_over_notnull[rename_key("originalgrade", "ZH1-")])
+
+mu, std = norm.fit(ZH1_over_notnull[rename_key("originalgrade", "ZH1-")])
+plt.plot(x_axis, norm.pdf(x_axis, mu, std)*count, label="Original ZH1 N", color="red")
+
+density = [sum(counts[0][0:6])]
+density.extend(counts[0][6:])
+density = np.array(density)
+limits = [14, 16, 18, 20, 22, 24, 26, 28, 30, 32, 34, 36, 38, 40]
+expected = norm.cdf(limits, mu, std) * float(count)
+propability = [expected[0]]
+propability.extend(expected[1:] - expected[0:-1])
+propability.append(count - expected[-1])
+
+test_stat, p_value = stats.chisquare(density, propability)
+# chi square test statistic and p value 
+print('chi_square_test_statistic is : ' +
+      str(test_stat)) 
+print('p_value : ' + str(p_value)) 
+print(stats.chi2.ppf(1-0.05, df=14)) 
+
+mu, std = norm.fit(ZH1_over_notnull[rename_key("revisedgrade", "ZH1-")])
+plt.plot(x_axis, norm.pdf(x_axis, mu, std)*count, label="Revised ZH1 N", color="lightcoral")
+mu, std = norm.fit(ZH2_over_notnull[rename_key("originalgrade", "ZH2-")])
+plt.plot(x_axis, norm.pdf(x_axis, mu, std), label="Original ZH2 N", color="blue")
+mu, std = norm.fit(ZH2_over_notnull[rename_key("revisedgrade", "ZH2-")])
+plt.plot(x_axis, norm.pdf(x_axis, mu, std), label="Revised ZH2 N", color="cornflowerblue")
+plt.xlabel('Point')
+plt.ylabel('Density')
+plt.legend()
+plt.show()
+
+
+
+coef = np.polyfit(ZH1_over_notnull[rename_key("originalgrade", "ZH1-")], ZH1_over_notnull[rename_key("revisedgrade", "ZH1-")], 1)
+lin = np.poly1d(coef)
+minmax = [0, 40] #[ZH1_over_notnull[rename_key("originalgrade", "ZH1-")].min(), ZH1_over_notnull[rename_key("originalgrade", "ZH1-")].max()]
+plt.plot(ZH1_over_notnull[rename_key("originalgrade", "ZH1-")], ZH1_over_notnull[rename_key("revisedgrade", "ZH1-")], '+', label="ZH1", color = "red")
+plt.plot(minmax, lin(minmax), ':', color="lightcoral", label="ZH1 lin")
+
+coef = np.polyfit(ZH2_over_notnull[rename_key("originalgrade", "ZH2-")], ZH2_over_notnull[rename_key("revisedgrade", "ZH2-")], 1)
+lin = np.poly1d(coef)
+minmax = [0, 40] #[ZH2_over_notnull[rename_key("originalgrade", "ZH2-")].min(), ZH2_over_notnull[rename_key("originalgrade", "ZH2-")].max()]
+plt.plot(ZH2_over_notnull[rename_key("originalgrade", "ZH2-")], ZH2_over_notnull[rename_key("revisedgrade", "ZH2-")], '*', label="ZH2", color = "blue")
+plt.plot(minmax, lin(minmax), '--', label="ZH2 lin", color="cornflowerblue")
+plt.axline((0,0), slope=1, c="black", label="y=x")
+plt.xlabel("Original grade")
+plt.ylabel("Revised grade")
+plt.legend()
+plt.savefig("revised_lin.svg", format = 'svg', dpi=300)
+plt.show()
+
+MSE_ZH1 = np.sqrt(mean_squared_error(ZH1_over_notnull[rename_key("originalgrade", "ZH1-")], ZH1_over_notnull[rename_key("revisedgrade", "ZH1-")]))
+print(MSE_ZH1)
+MSE_ZH2 = np.sqrt(mean_squared_error(ZH2_over_notnull[rename_key("originalgrade", "ZH2-")], ZH2_over_notnull[rename_key("revisedgrade", "ZH2-")]))
+print(MSE_ZH2)
+
+plt.boxplot([ZH1_over_notnull["ZH1-originalgrade"], ZH1_over_notnull["ZH1-revisedgrade"], ZH2_over_notnull["ZH2-originalgrade"], ZH2_over_notnull["ZH2-revisedgrade"]], 
+            labels=["Original ZH1", "Revised ZH1", "Original ZH2", "Revised ZH2"])
+plt.title('Box Plot of grades')
+plt.ylabel('Values')
+plt.show()
+
+plt.boxplot([df[(df[rename_key("overridden", "ZH1-")] == "Yes")][rename_key("diff", "ZH1-")], 
+             df[(df[rename_key("overridden", "ZH2-")] == "Yes")][rename_key("diff", "ZH2-")]], labels=["Only difference ZH1", "Only difference ZH2"])
+plt.title('Box Plot of grades')
+plt.ylabel('Values')
+plt.show()
+
+diff_by_graders_zh1 = [df[(df[rename_key("grader", "ZH1-")] == g)][rename_key("diff", "ZH1-")] for g in graders.keys()]
+diff_by_graders_zh2 = [df[(df[rename_key("grader", "ZH2-")] == g)][rename_key("diff", "ZH2-")] for g in graders.keys()]
+diff_by_graders_pzh1 = [df[(df[rename_key("grader", "PZH1-")] == g)][rename_key("diff", "PZH1-")] for g in graders.keys()]
+diff_by_graders_pzh2 = [df[(df[rename_key("grader", "PZH2-")] == g)][rename_key("diff", "PZH2-")] for g in graders.keys()]
+
+graders_anonym = [chr(i) for i in range(ord('A'),ord('A')+len(graders))]
+fig, ax1 = plt.subplots()
+
+ax1.set_ylabel("Corrected piece (bar)")
+ax1.bar(range(1, len(graders_anonym) + 1), [s.size for s in diff_by_graders_zh1], alpha=0.3)
+ax1.set_xlabel("Grader")
+
+color = 'tab:red'
+ax2 = ax1.twinx()
+ax2.set_ylabel("The points improved (boxplot)")
+ax2.boxplot(diff_by_graders_zh1, labels=graders_anonym)
+
+fig.tight_layout() 
+plt.savefig("graders_ZH1.svg", format = 'svg', dpi=300)
+plt.show()
+
+fig, ax1 = plt.subplots()
+ax1.set_ylabel("Corrected piece (bar)")
+ax1.bar(range(1, len(graders_anonym) + 1), [s.size for s in diff_by_graders_zh2], alpha=0.3)
+ax1.set_xlabel("Grader")
+
+color = 'tab:red'
+ax2 = ax1.twinx()
+ax2.set_ylabel("The points improved (boxplot)")
+ax2.boxplot(diff_by_graders_zh2, labels=graders_anonym)
+
+fig.tight_layout() 
+plt.savefig("graders_ZH2.svg", format = 'svg', dpi=300)
+plt.show()
+
+
+diff_summary = [pd.concat([diff_by_graders_zh1[i], diff_by_graders_zh2[i], diff_by_graders_pzh1[i], diff_by_graders_pzh2[i]]) for i in range(0, len(diff_by_graders_zh1))]
+plt.boxplot(diff_summary, labels=graders_anonym)
+plt.title('Box Plot of difference by graders')
+plt.ylabel('Values')
+plt.show()
+
+diff_by_graders_zh1 = [df[(df[rename_key("grader", "ZH1-")] == g)][rename_key("originalgrade", "ZH1-")] for g in graders.keys()]
+diff_by_graders_zh2 = [df[(df[rename_key("grader", "ZH2-")] == g)][rename_key("originalgrade", "ZH2-")] for g in graders.keys()]
+diff_by_graders_pzh1 = [df[(df[rename_key("grader", "PZH1-")] == g)][rename_key("originalgrade", "PZH1-")] for g in graders.keys()]
+diff_by_graders_pzh2 = [df[(df[rename_key("grader", "PZH2-")] == g)][rename_key("originalgrade", "PZH2-")] for g in graders.keys()]
+
+graders_anonym = [chr(i) for i in range(ord('A'),ord('A')+len(graders))]
+
+plt.boxplot(diff_by_graders_zh1, labels=graders_anonym)
+plt.title('Box Plot of teachers ZH1')
+plt.ylabel('Original grade')
+plt.show()
+
+
+plt.boxplot(diff_by_graders_zh2, labels=graders_anonym)
+plt.title('Box Plot of teachers ZH2')
+plt.ylabel('Original grade')
+plt.show()
\ No newline at end of file