chuyển đổi cấu trúc

2026-07-08 15:11:34 +07:00
parent b041ccc11a
commit b22d327c5f
45 changed files with 201 additions and 4 deletions
@@ -0,0 +1,8 @@
+,Temperature,Noise,Stormwind,Respiratory,Exercises,Culture,Beauty,Education,Society,Spirit,Dirty,Unsafe,Danger
+Park,0.5861277411758214,0.46872135461497316,0.5433785341732876,0.552634098528652,0.6278282242144403,0.5249311710036608,0.6010523214572643,0.6698801970338424,0.592175252811864,0.5873765414225279,-0.4119222589542689,-0.36935992311052424,-0.47499404654301275
+Residential,0.610813681241285,0.5010876976294922,0.5812495058768619,0.6308055529227076,0.6905711940417876,0.5336682510392587,0.7050811390341176,0.6880894492038738,0.6281007540295365,0.6725382465666585,-0.4172763562532894,-0.3683423748854766,-0.47006223532477215
+Garden,0.45224716780106355,0.3994427050538716,0.41423108867853403,0.4171592620121273,0.5703375228779752,0.42420205471782707,0.5225523665124601,0.5586006221718542,0.4604821197705184,0.4741367465117808,-0.3021035545730409,-0.2690208910002917,-0.33928178580122825
+Rooftop,0.4960914195506356,0.4389763334105073,0.4741573996062278,0.4043299683546717,0.5234138548608293,0.40036627443478834,0.5613291264190557,0.5137072417522562,0.4507351073588117,0.533245963504779,-0.3044536381978058,-0.3287078138585324,-0.3540628872780644
+Recreation,0.40870892534299835,0.36018022246155923,0.38726035929996405,0.327640261869753,0.4435840120247182,0.41597785467253595,0.41702432907643733,0.46845721173859056,0.337735516408859,0.346506311430844,-0.08584770592904668,-0.14113338302330178,-0.2698727403572271
+Agriculture,-0.3186847619885256,-0.22410153490737583,-0.31116783542012505,-0.31164843416963245,-0.29672463363412205,-0.08295065998935172,-0.407725512606722,-0.34041202068253673,-0.33774567405476646,-0.3661505107268029,0.4906684665223075,0.4667053886536356,0.5249948486680992
+Nature,-0.4377274782376116,-0.2893202018289482,-0.47487184947631206,-0.4416794342162952,-0.4395911413850507,-0.2583285861208512,-0.4891875676625891,-0.4669790589031381,-0.4642512201471137,-0.5258378364708906,0.5182221077803194,0.5103204875306926,0.5338960402111311
@@ -0,0 +1,14 @@
+,0,1
+Temperature,-0.15445511914940652,0.006990515901429791
+Noise,-0.07615894298740968,0.00700492282153885
+Stormwind,-0.1563914823055307,0.007878888704259901
+Respiratory,-0.14800873993854352,-0.011635761620414456
+Exercises,-0.16208617036601874,-0.0003806098068870642
+Culture,-0.04550439013316019,0.00881620114706321
+Beauty,-0.19796961667785237,-0.0012069882779817856
+Education,-0.17984311395007466,0.008345565782577814
+Society,-0.16523244764182207,-0.011179747262609863
+Spirit,-0.1938278638629291,-0.01479224672248243
+Dirty,0.662990083180058,0.026367956173616355
+Unsafe,0.6446258846820032,-0.0004123561994046179
+Danger,0.7399178159934315,-0.028168947040193765
@@ -0,0 +1,8 @@
+,0,1
+Park,-0.20737329331501148,-0.002436125971632047
+Residential,-0.2155834258211865,-0.011730849372120936
+Garden,-0.14244079709884974,-0.004395477787898166
+Rooftop,-0.15500847340894128,-0.003399988119557015
+Recreation,-0.0682883185805637,0.027918075759378963
+Agriculture,0.5601149289877119,-0.005751109303690985
+Nature,0.6998145216538895,-0.0007509375795379177
@@ -0,0 +1,8 @@
+,Temperature,Noise,Stormwind,Respiratory,Exercises,Culture,Beauty,Education,Society,Spirit,Dirty,Unsafe,Danger
+Park,0.5861277411758214,0.4687213546149731,0.5433785341732876,0.552634098528652,0.6278282242144403,0.5249311710036608,0.6010523214572643,0.6698801970338424,0.592175252811864,0.5873765414225279,-0.4119222589542689,-0.3693599231105242,-0.4749940465430127
+Residential,0.610813681241285,0.5010876976294922,0.5812495058768619,0.6308055529227076,0.6905711940417876,0.5336682510392587,0.7050811390341176,0.6880894492038738,0.6281007540295365,0.6725382465666585,-0.4172763562532894,-0.3683423748854766,-0.4700622353247721
+Garden,0.4522471678010635,0.3994427050538716,0.414231088678534,0.4171592620121273,0.5703375228779752,0.424202054717827,0.5225523665124601,0.5586006221718542,0.4604821197705184,0.4741367465117808,-0.3021035545730409,-0.2690208910002917,-0.3392817858012282
+Rooftop,0.4960914195506356,0.4389763334105073,0.4741573996062278,0.4043299683546717,0.5234138548608293,0.4003662744347883,0.5613291264190557,0.5137072417522562,0.4507351073588117,0.533245963504779,-0.3044536381978058,-0.3287078138585324,-0.3540628872780644
+Recreation,0.4087089253429983,0.3601802224615592,0.387260359299964,0.327640261869753,0.4435840120247182,0.4159778546725359,0.4170243290764373,0.4684572117385905,0.337735516408859,0.346506311430844,-0.0858477059290466,-0.1411333830233017,-0.2698727403572271
+Agriculture,-0.3186847619885256,-0.2241015349073758,-0.311167835420125,-0.3116484341696324,-0.296724633634122,-0.0829506599893517,-0.407725512606722,-0.3404120206825367,-0.3377456740547664,-0.3661505107268029,0.4906684665223075,0.4667053886536356,0.5249948486680992
+Nature,-0.4377274782376116,-0.2893202018289482,-0.474871849476312,-0.4416794342162952,-0.4395911413850507,-0.2583285861208512,-0.4891875676625891,-0.4669790589031381,-0.4642512201471137,-0.5258378364708906,0.5182221077803194,0.5103204875306926,0.5338960402111311
@@ -0,0 +1,93 @@
+import pandas as pd
+import numpy as np
+import matplotlib.pyplot as plt
+import prince
+
+# Load data
+df = pd.read_excel('Data_VN_filter_v5.xlsx')
+
+# Extract variables
+ugs_cols = ['Park', 'Residential', 'Garden', 'Rooftop', 'Recreation', 'Agriculture', 'Nature']
+
+# Detailed ESS and DES variables
+ess_des_detailed = [
+    # RES (Điều hòa)
+    'Temperature', 'Noise', 'Stormwind', 'Respiratory',
+    # CES (Văn hóa)
+    'Exercises', 'Culture', 'Beauty', 'Education', 'Society', 'Spirit',
+    # DES (Bất lợi)
+    'Dirty', 'Unsafe', 'Danger'
+]
+
+# Ensure the columns exist in df
+available_detailed = [col for col in ess_des_detailed if col in df.columns]
+
+# We want to run CA on the relationships between the 7 UGS types and the 13 specific services.
+# Dữ liệu thang đo Likert thường là dữ liệu thứ bậc (Ordinal), nên dùng Spearman's rho sẽ chính xác hơn Pearson.
+data = df[ugs_cols + available_detailed].dropna()
+corr = data.corr(method='spearman').loc[ugs_cols, available_detailed]
+
+# Tương tự như trước, dịch chuyển ma trận hệ số RHO để các giá trị đều dương (dùng cho CA)
+# Alternatively, since prince CA handles frequencies, we can just pass the raw data? 
+# If we pass raw data of shape (307, 20), CA will treat rows as individuals. 
+# We want to see relationships between UGS and Detailed Services. Passing the correlation matrix (shifted) is a good proxy for similarity.
+corr_shifted = corr + 1
+
+# Initialize CA
+ca = prince.CA(n_components=2, n_iter=3, copy=True, check_input=True, engine='scipy', random_state=42)
+ca = ca.fit(corr_shifted)
+
+# Extract column and row coordinates
+row_coords = ca.row_coordinates(corr_shifted) # UGS
+col_coords = ca.column_coordinates(corr_shifted) # Detailed Services
+
+# Save to CSV
+corr.to_csv('CA_Detailed_Correlation_Matrix.csv')
+row_coords.to_csv('CA_Detailed_UGS_Coords.csv')
+col_coords.to_csv('CA_Detailed_Services_Coords.csv')
+
+# Plot Biplot
+fig, ax = plt.subplots(figsize=(20, 15))
+
+# Plot UGS points (blue dots)
+p_ugs = ax.scatter(row_coords[0], row_coords[1], c='blue', label='UGS Types', s=80, marker='o', edgecolors='black')
+
+# Plot ESS/DES points (red/green squares)
+res_cols = ['Temperature', 'Noise', 'Stormwind', 'Respiratory']
+ces_cols = ['Exercises', 'Culture', 'Beauty', 'Education', 'Society', 'Spirit']
+des_cols = ['Dirty', 'Unsafe', 'Danger']
+
+p_ess = []
+for col in available_detailed:
+    color = 'green' if col in res_cols else ('purple' if col in ces_cols else 'red')
+    p = ax.scatter(col_coords.loc[col, 0], col_coords.loc[col, 1], c=color, marker='s', s=80, edgecolors='black')
+    p_ess.append(p)
+
+# Gắn nhãn (Text) tĩnh để đảm bảo tính nhất quán (Deterministic) 100% mỗi lần chạy
+# Phân tách vị trí nhãn: UGS nằm lệch trên-trái, ESS/DES nằm lệch dưới-phải
+for i, txt in enumerate(ugs_cols):
+    ax.annotate(txt, (row_coords.iloc[i, 0], row_coords.iloc[i, 1]), 
+                xytext=(-10, 10), textcoords='offset points', 
+                color='blue', fontweight='bold', fontsize=12, ha='right', va='bottom')
+
+for col in available_detailed:
+    color = 'green' if col in res_cols else ('purple' if col in ces_cols else 'red')
+    ax.annotate(col, (col_coords.loc[col, 0], col_coords.loc[col, 1]), 
+                xytext=(10, -10), textcoords='offset points', 
+                color=color, fontsize=12, fontweight='bold', ha='left', va='top')
+
+# Add dummy plots for legend
+ax.scatter([], [], c='green', marker='s', label='RES (Điều hòa)')
+ax.scatter([], [], c='purple', marker='s', label='CES (Văn hóa)')
+ax.scatter([], [], c='red', marker='s', label='DES (Bất lợi)')
+
+ax.axhline(0, color='grey', linestyle='--', linewidth=1)
+ax.axvline(0, color='grey', linestyle='--', linewidth=1)
+ax.set_title('Detailed CA Biplot (UGS vs Specific ESS/DES)', fontsize=16)
+ax.set_xlabel('Component 0', fontsize=12)
+ax.set_ylabel('Component 1', fontsize=12)
+ax.legend(loc='center left', bbox_to_anchor=(1.02, 0.5), fontsize=12)
+plt.grid(True, linestyle=':', alpha=0.6)
+plt.savefig('CA_plot_detailed.png', dpi=300, bbox_inches='tight')
+print("\nDetailed CA Plot saved to CA_plot_detailed.png")
+print("Detailed data exported to CSV files.")
@@ -0,0 +1,24 @@
+import pandas as pd
+import seaborn as sns
+import matplotlib.pyplot as plt
+
+# Đọc ma trận RHO đã tính từ file CSV
+corr = pd.read_csv('CA_Detailed_Correlation_Matrix.csv', index_col=0)
+
+# Vẽ biểu đồ nhiệt (Heatmap)
+plt.figure(figsize=(14, 8))
+sns.heatmap(corr, annot=True, fmt=".2f", cmap="coolwarm", center=0, 
+            vmin=-1, vmax=1, linewidths=.5, cbar_kws={"shrink": .8})
+
+plt.title('Spearman RHO Correlation Matrix (UGS vs Specific ESS/DES)', fontsize=16, pad=20)
+plt.ylabel('UGS Types', fontsize=12)
+plt.xlabel('ESS/DES Variables', fontsize=12)
+
+# Xoay nhãn trục X để dễ đọc hơn
+plt.xticks(rotation=45, ha='right')
+plt.yticks(rotation=0)
+
+plt.tight_layout()
+plt.savefig('RHO_Heatmap.png', dpi=300)
+corr.to_csv('RHO_Matrix.csv')
+print("Heatmap saved to RHO_Heatmap.png and data to RHO_Matrix.csv")