for benefit cost, calculate travel time instead

scripts/summarize/benefit_cost.py

@@ -10,13 +10,13 @@
 # this should be a run or config parameter
 model_dir = 'C:/soundcast615/'
 
-h5_base_file = 'outputs/daysim_outputs3.h5'
+h5_base_file = 'outputs/daysim_outputs.h5'
 h5_base_name = 'Base'
-h5_scen_file = 'outputs/daysim_outputs1.h5'
+h5_scen_file = 'outputs/daysim_outputs2040luv.h5'
 h5_scen_name = 'Scenario'
-report_output_location = 'outputs/benefits_zone.csv'
+report_output_location = 'outputs/time_zone.csv'
 guidefile = 'scripts/summarize/CatVarDict.xlsx'
-zone_dim = 4050 #hard code should be in the config
+zone_dim = 4000 #hard code should be in the config
 #output
 
 ##################################
@@ -41,60 +41,61 @@
 dtaz_base = np.asarray(trips_base["dtaz"])
 dtaz_base = dtaz_base.astype('int')
 
-vot_base = np.asarray(trips_base["vot"])
+#vot_base = np.asarray(trips_base["vot"])
 time_base = np.asarray(trips_base["travtime"])
-cost_base =vot_base*time_base
+#cost_base =vot_base*time_base
 
 otaz_scen = np.asarray(trips_scen["otaz"])
 otaz_scen = otaz_scen.astype('int')
 
 dtaz_scen = np.asarray(trips_scen["dtaz"])
-dtaz_scen = dtaz_base.astype('int')
+dtaz_scen = dtaz_scen.astype('int')
 
-vot_scen = np.asarray(trips_scen["vot"])
+#vot_scen = np.asarray(trips_scen["vot"])
 time_scen = np.asarray(trips_scen["travtime"])
-cost_scen =vot_scen*time_scen
+#cost_scen =vot_scen*time_scen
 
 
 trip_base_matrix = np.zeros((zone_dim,zone_dim), np.float64)
-cost_base_matrix= np.zeros((zone_dim,zone_dim), np.float64)
+time_base_matrix= np.zeros((zone_dim,zone_dim), np.float64)
 trip_scen_matrix = np.zeros((zone_dim,zone_dim), np.float64)
-cost_scen_matrix= np.zeros((zone_dim,zone_dim), np.float64)
+time_scen_matrix= np.zeros((zone_dim,zone_dim), np.float64)
 
 # Fill up the base matrices from the arrays of data because they will be faster to work with
-for trip in range (0, len(otaz_base)):
+for trip in range (0, len(otaz_base)-1):
     this_otaz = otaz_base[trip]
     this_dtaz = dtaz_base[trip]
-    this_cost = cost_base[trip]
+    this_time = time_base[trip]
     trip_base_matrix [this_otaz, this_dtaz] = trip_base_matrix[this_otaz, this_dtaz] + 1
-    cost_base_matrix [this_otaz, this_dtaz] = cost_base_matrix[this_otaz, this_dtaz] + this_cost
+    time_base_matrix [this_otaz, this_dtaz] = time_base_matrix[this_otaz, this_dtaz] + this_time
 
 # Fill up the scenario matrices
-for trip in range (0, len(otaz_scen)):
-    this_otaz = otaz_scen[trip]
-    this_dtaz = dtaz_scen[trip]
-    this_cost = cost_scen[trip]
+for trip_s in range (0, len(otaz_scen)-1):
+    this_otaz = otaz_scen[trip_s]
+    this_dtaz = dtaz_scen[trip_s]
+    this_time = time_scen[trip_s]
     trip_scen_matrix [this_otaz, this_dtaz] = trip_scen_matrix[this_otaz, this_dtaz] + 1
-    cost_scen_matrix [this_otaz, this_dtaz] = cost_scen_matrix[this_otaz, this_dtaz] + this_cost
+    time_scen_matrix [this_otaz, this_dtaz] = time_scen_matrix[this_otaz, this_dtaz] + this_time
 
-cost_diff = (cost_scen_matrix - cost_base_matrix)
-trips_total = (trip_scen_matrix + trip_base_matrix)/2
+time_diff = (time_scen_matrix - time_base_matrix)
+trips_total = (trip_scen_matrix + trip_base_matrix)
 
-benefit_matrix = cost_diff*trips_total
+avg_time_diff_matrix = time_diff/trips_total
+avg_time_diff_matrix[np.isnan(avg_time_diff_matrix)] =0
 
 taz_labels = np.array(range(1, zone_dim + 1))
 
 #benefits by origin:
-bens_by_origin = np.nansum(benefit_matrix , axis = 1, dtype=np.float64)
+time_by_origin = np.mean(avg_time_diff_matrix, axis = 1, dtype=np.float64)
 
 #benefits by destination:
-bens_by_destination = np.nansum(benefit_matrix, axis = 0, dtype=np.float64)
+time_by_destination = np.mean(avg_time_diff_matrix, axis = 0, dtype=np.float64)
 
 #combine arrays:
-all_colls = zip(taz_labels,bens_by_origin, bens_by_destination)
+all_colls = zip(taz_labels,time_by_origin, time_by_destination)
 
 #convert to dataframe
-results = pd.DataFrame(data=all_colls, columns=['taz','origin_benefit', 'destination_benefit'])
+results = pd.DataFrame(data=all_colls, columns=['taz','origin_time_diff', 'destination_time_diff'])
 
 results.to_csv(report_output_location)