Skip to content

som_tools

Methods to preprocess input data for SOM.

get_expert_ids(df)

Returns list of expert id column names from dataframe using regex.

Parameters:

Name Type Description Default
df DataFrame

dataframe containing expert idc columns.

 required

Returns:

Type Description
list

list
Source code in src/som_tools.py 
11
12
13
14
15
16
17
18
19
20
21
def get_expert_ids(df: pd.DataFrame) -> list:
    '''
    Returns list of expert id column names from dataframe using regex.

    Arguments:
        df (DataFrame): dataframe containing expert idc columns.

    Returns:
        list
    '''
    return df.filter(regex='^(100|[1-9]?[0-9])$').columns

process_input_data(config)

Reads in data and processes raw input data to usable form.

Parameters:

Name Type Description Default
config dict

dictionary loaded from configuration file.

 required

Returns:

Name Type Description
measure_survey_df DataFrame

contains the measure survey data of expert panels.
pressure_survey_df DataFrame

contains the pressure survey data of expert panels.
data dict

container for general data dataframes:

  • measure (DataFrame):

    • ID: unique measure identifier.
    • measure: name / description column.

  • activity (DataFrame):

    • ID: unique activity identifier.
    • activity: name / description column.

  • pressure (DataFrame):

    • ID: unique pressure identifier.
    • pressure: name / description column.

  • state (DataFrame):

    • ID: unique state identifier.
    • state: name / description column.

  • area (DataFrame):

    • ID: unique area identifier.
    • area: name / description column.

  • measure_effects (DataFrame): measure effects on activities, pressures, states.

  • pressure_contributions (DataFrame): pressure contributions to states.
  • thresholds (DataFrame): changes in states required to meet specific target thresholds.
  • cases (DataFrame): measure implementations in areas.
  • activity_contributions (DataFrame): activity contributions to pressures.
  • overlaps (DataFrame): measure-measure interactions.
  • development_scenarios (DataFrame): changes in human activities.
  • subpressures (DataFrame): pressure-pressure interactions.
Source code in src/som_tools.py 
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
def process_input_data(config: dict) -> dict[str, pd.DataFrame | dict[str, pd.DataFrame]]:
    """
    Reads in data and processes raw input data to usable form.

    Arguments:
        config (dict): dictionary loaded from configuration file.

    Returns:
        measure_survey_df (DataFrame): contains the measure survey data of expert panels.
        pressure_survey_df (DataFrame): contains the pressure survey data of expert panels.
        data (dict): container for general data dataframes:

            - measure (DataFrame):
                - ID: unique measure identifier.
                - measure: name / description column.

            - activity (DataFrame):
                - ID: unique activity identifier.
                - activity: name / description column.

            - pressure (DataFrame):
                - ID: unique pressure identifier.
                - pressure: name / description column.

            - state (DataFrame):
                - ID: unique state identifier.
                - state: name / description column.

            - area (DataFrame):
                - ID: unique area identifier.
                - area: name / description column.

            - measure_effects (DataFrame): measure effects on activities, pressures, states.
            - pressure_contributions (DataFrame): pressure contributions to states.
            - thresholds (DataFrame): changes in states required to meet specific target thresholds.
            - cases (DataFrame): measure implementations in areas.
            - activity_contributions (DataFrame): activity contributions to pressures.
            - overlaps (DataFrame): measure-measure interactions.
            - development_scenarios (DataFrame): changes in human activities.
            - subpressures (DataFrame): pressure-pressure interactions.
    """
    #
    # measure survey data
    #

    file_name = os.path.realpath(config['input_data_legacy']['measure_effect_input'])
    if not os.path.isfile(file_name): file_name = os.path.join(os.path.dirname(os.path.realpath(__file__)), config['input_data_legacy']['measure_effect_input'])
    measure_effects = process_measure_survey_data(file_name)

    #
    # pressure survey data (combined pressure contributions and GES threshold)
    #

    file_name = os.path.realpath(config['input_data_legacy']['pressure_state_input'])
    if not os.path.isfile(file_name): file_name = os.path.join(os.path.dirname(os.path.realpath(__file__)), config['input_data_legacy']['pressure_state_input'])
    pressure_contributions, thresholds = process_pressure_survey_data(file_name)

    #
    # measure / pressure / activity / state links
    #

    # read core object descriptions
    # i.e. ids for measures, activities, pressures and states
    file_name = os.path.realpath(config['input_data_legacy']['general_input'])
    if not os.path.isfile(file_name): file_name = os.path.join(os.path.dirname(os.path.realpath(__file__)), config['input_data_legacy']['general_input'])
    id_sheets = config['input_data_legacy']['general_input_sheets']['ID']
    data = read_ids(file_name=file_name, id_sheets=id_sheets)

    #
    # read case input
    #

    sheet_name = config['input_data_legacy']['general_input_sheets']['case']
    cases = read_cases(file_name=file_name, sheet_name=sheet_name)

    #
    # read activity contribution data
    #

    sheet_name = config['input_data_legacy']['general_input_sheets']['postprocess']
    activity_contributions = read_activity_contributions(file_name=file_name, sheet_name=sheet_name)

    #
    # read overlap data
    #

    sheet_name = config['input_data_legacy']['general_input_sheets']['overlaps']
    overlaps = read_overlaps(file_name=file_name, sheet_name=sheet_name)

    #
    # read activity development scenario data
    #

    sheet_name = config['input_data_legacy']['general_input_sheets']['development_scenarios']
    development_scenarios = read_development_scenarios(file_name=file_name, sheet_name=sheet_name)

    #
    # read subpressures links
    #

    sheet_name = config['input_data_legacy']['general_input_sheets']['subpressures']
    subpressures = read_subpressures(file_name=file_name, sheet_name=sheet_name)

    data.update({
        'measure_effects': measure_effects, 
        'pressure_contributions': pressure_contributions, 
        'thresholds': thresholds, 
        'cases': cases, 
        'activity_contributions': activity_contributions, 
        'overlaps': overlaps, 
        'development_scenarios': development_scenarios, 
        'subpressures': subpressures
    })

    return data

process_measure_survey_data(file_name)

This method reads input from the excel file containing data about measure reduction efficiencies on activities, pressures and states.

Parameters:

Name Type Description Default
file_name str

path of survey excel file.

 required

Returns:

Name Type Description
survey_df DataFrame

processed survey data information:

  • measure (int): measure id.
  • activity (int): activity id.
  • pressure (int): pressure id.
  • state (int): state id (if defined, [nan] if no state).
  • reduction (list[float]): probability distribution represented as list.
Source code in src/som_tools.py 
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
def process_measure_survey_data(file_name: str) -> pd.DataFrame:
    """
    This method reads input from the excel file containing data about measure reduction efficiencies 
    on activities, pressures and states.

    Arguments:
        file_name (str): path of survey excel file.

    Returns:
        survey_df (DataFrame): processed survey data information:

            - measure (int): measure id.
            - activity (int): activity id.
            - pressure (int): pressure id.
            - state (int): state id (if defined, [nan] if no state).
            - reduction (list[float]): probability distribution represented as list.
    """
    #
    # read information sheet from input Excel file
    #

    data = pd.read_excel(io=file_name, sheet_name=None, header=None)
    sheet_names = list(data.keys())

    mteq = data[sheet_names[0]]
    mteq.columns = mteq.iloc[0].values
    mteq = mteq[1:]

    measure_survey_data = {}
    for id in range(1, len(sheet_names)):
        measure_survey_data[id] = data[sheet_names[id]]

    #
    # preprocess values
    #

    mteq.loc[:, 'State'] = [x.split(';') if type(x) == str else x for x in mteq['State']]

    #
    # create new dataframe
    #

    cols = ['survey_id', 'title', 'block', 'measure', 'activity', 'pressure', 'state']
    survey_df = pd.DataFrame(columns=cols)

    block_number = 0    # represents the survey block

    # for every survey sheet
    for survey_id in measure_survey_data:

        survey_info = mteq[mteq['Survey ID'] == survey_id]  # select the rows linked to current survey

        end = 0     # represents last column index of the question set
        for row, amt in enumerate(survey_info['AMT']):  # for each set of questions (row in MTEQ)

            end = end + (2 * amt + 1)     # end column index for data
            start = end - (2 * amt)     # start column index for data

            # create list to describe the data on each row
            titles = ['expected value', 'variance'] * amt
            titles.append('max effectiveness')
            titles.append('expert weights')

            # select current question column names as measure ids
            measures = measure_survey_data[survey_id].iloc[0, start:end].tolist()
            measures.append(np.nan)
            measures.append(np.nan)

            # create lists to hold ids and format each row as a list
            ids = {}
            for category in ['Activity', 'Pressure', 'State']:
                category_ids = survey_info[category].iloc[row]
                if isinstance(category_ids, str):
                    ids[category] = [[int(x) for x in category_ids.split(';') if x != '']] * amt * 2
                elif isinstance(category_ids, list):
                    ids[category] = [[int(x) for x in category_ids if x != '']] * amt * 2
                elif isinstance(category_ids, float) or isinstance(category_ids, int):
                    ids[category] = [[category_ids if not np.isnan(category_ids) else np.nan]] * amt * 2
                else:
                    ids[category] = [category_ids] * amt * 2
                ids[category].append(np.nan)
                ids[category].append(np.nan)

            # in MTEQ sheet, find all expert weight columns, get the values for the current row, set empty cells to 1
            expert_cols = [True if 'exp' in col.lower() else False for col in survey_info.columns]
            expert_weights = survey_info.loc[:, expert_cols].iloc[row]
            expert_weights = expert_weights.astype(float).fillna(1).astype(int) # convert to float first so fillna() works without warning

            data = measure_survey_data[survey_id].loc[1:, start:end]    # select current question answers
            data[end+1] = expert_weights  # create column for expert weights
            for expert, weight in enumerate(expert_weights, 1): # for each row (expert) in weights
                data.loc[expert, end+1] = weight    # set the weight as the value
            data = data.transpose() # transpose so that experts are columns and measures are rows

            # add survey info to each entry in the data 
            data['survey_id'] = [survey_id] * len(data) # new column with survey_id for every row
            data['title'] = titles
            data['block'] = [block_number] * len(data)  # new column with block_number for every row
            data['measure'] = measures
            data['activity'] = ids['Activity']
            data['pressure'] = ids['Pressure']
            data['state'] = ids['State']

            with warnings.catch_warnings(action='ignore'):
                survey_df = pd.concat([survey_df, data], ignore_index=True, sort=False)
            block_number = block_number + 1

    # select column names corresponding to expert ids (any number between 1 and 100)
    expert_ids = get_expert_ids(survey_df)

    #
    # Adjust answers by scaling factor
    #

    block_ids = survey_df.loc[:,'block'].unique()   # find unique block ids
    for b_id in block_ids:  # for each block
        block = survey_df.loc[survey_df['block'] == b_id, :]    # select all rows with current block id
        for col in block:   # for each column
            if isinstance(col, int):    # if it is an expert answer
                # from the column, select the expected values and variances
                expected_value = block.loc[block['title']=='expected value', col]
                variance = block.loc[block['title']=='variance', col]
                # skip if no questions were answered
                if expected_value.isnull().all():
                    block.loc[block['title']=='variance', col] = np.nan     # also set all variances to null
                    continue
                if variance.isnull().all():
                    block.loc[block['title']=='expected value', col] = np.nan     # also set all expected values to null
                    continue
                # find the highest value of the answers
                max_expected_value = expected_value.max()
                # find the max effectiveness estimated by the expert
                max_effectiveness = block.loc[block['title']=='max effectiveness', col].values[0]
                # calculate scaling factor
                if np.isnan(max_effectiveness):
                    # set all values to null if no max effectiveness (in column, for current block)
                    survey_df.loc[survey_df['block'] == b_id, col] = np.nan
                elif max_effectiveness == 0 or max_expected_value == 0:
                    # scale all expected values to 0 if max effectiveness is zero or all expected values are zero
                    survey_df.loc[(survey_df['block'] == b_id) & (survey_df['title'] == 'expected value'), col] = 0
                else:
                    # get the scaling factor
                    scaling_factor = np.divide(max_expected_value, max_effectiveness)
                    # divide the expected values by the new scaling factor
                    survey_df.loc[(survey_df['block'] == b_id) & (survey_df['title'] == 'expected value'), col] = np.divide(expected_value, scaling_factor)

    #
    # Calculate effectiveness range boundaries
    #

    # create new rows for 'effectiveness lower' and 'effectiveness upper' bounds after variance rows
    new_rows = []
    for i, row in survey_df.iterrows():
        new_rows.append(row)
        if row['title'] == 'variance':
            # create lower bound
            min_row = survey_df.loc[i].copy()
            min_row['title'] = 'effectiveness lower'
            min_row[expert_ids] = np.nan
            new_rows.append(min_row)
            # create upper bound
            max_row = survey_df.loc[i].copy()
            max_row['title'] = 'effectiveness upper'
            max_row[expert_ids] = np.nan
            new_rows.append(max_row)
    survey_df = pd.DataFrame(new_rows, columns=survey_df.columns)
    survey_df.reset_index(drop=True, inplace=True)
    # set values for 'effectiveness lower' and 'effectiveness upper' bounds rows
    # calculated as follows:
    #   lower boundary:
    #       if expected_value + variance / 2 > 100:
    #           boundary = 100 - variance
    #       else:
    #           if expected_value - variance / 2 < 0:
    #               boundary = 0
    #           else:
    #               boundary = expected_value - variance / 2
    #   upper boundary:
    #       if expected_value - variance / 2 < 0:
    #           boundary = variance
    #       else:
    #           if expected_value + variance / 2 > 100:
    #               boundary = 100
    #           else:
    #               boundary = expected_value + variance / 2
    for i, row in survey_df.iterrows():
        if row['title'] == 'effectiveness lower':
            expected_value = survey_df.iloc[i-2][expert_ids]
            variance = survey_df.iloc[i-1][expert_ids]
            reach_upper_limit = expected_value + variance / 2 > 100 # boolean array
            row_values = survey_df.loc[i, expert_ids]
            row_values[reach_upper_limit] = 100 - variance
            row_values[~reach_upper_limit] = expected_value - variance / 2
            row_values[row_values < 0] = 0
            survey_df.loc[i, expert_ids] = row_values
        if row['title'] == 'effectiveness upper':
            expected_value = survey_df.iloc[i-3][expert_ids]
            variance = survey_df.iloc[i-2][expert_ids]
            reach_lower_limit = expected_value - variance / 2 < 0   # boolean array
            row_values = survey_df.loc[i, expert_ids]
            row_values[reach_lower_limit] = variance
            row_values[~reach_lower_limit] = expected_value + variance / 2
            row_values[row_values > 100] = 100
            survey_df.loc[i, expert_ids] = row_values

    #
    # Calculate probability distributions
    #

    # add a new column for the probability
    survey_df['reduction'] = pd.Series([np.nan] * len(survey_df), dtype='object')

    # access expert answer columns, separate rows by type of answer
    expecteds = survey_df[expert_ids].loc[survey_df['title'] == 'expected value']
    lower_boundaries = survey_df[expert_ids].loc[survey_df['title'] == 'effectiveness lower']
    upper_boundaries = survey_df[expert_ids].loc[survey_df['title'] == 'effectiveness upper']
    weights = survey_df.loc[survey_df['title'] == 'expert weights', np.insert(expert_ids, 0, 'block')]
    blocks = survey_df['block'].loc[(survey_df['title'] == 'expected value')]
    # go through each measure-activity-pressure link
    for num in expecteds.index:
        # access current row data and convert to 1-D arrays
        b_id = blocks.loc[num]
        e = expecteds.loc[num].to_numpy().astype(float)
        l = lower_boundaries.loc[num+2].to_numpy().astype(float)
        u = upper_boundaries.loc[num+3].to_numpy().astype(float)
        w = weights.loc[weights['block'] == b_id, expert_ids].to_numpy().astype(float).flatten()
        # get expert probability distribution
        prob_dist = get_prob_dist(expecteds=e, 
                                  lower_boundaries=l, 
                                  upper_boundaries=u, 
                                  weights=w)

        survey_df.at[num, 'reduction'] = prob_dist

    #
    # Remove rows and columns that are not needed anymore
    #

    for title in ['max effectiveness', 'variance', 'effectiveness lower', 'effectiveness upper', 'expert weights']:
        survey_df = survey_df.loc[survey_df['title'] != title]
    survey_df = survey_df.drop(columns=expert_ids)
    survey_df = survey_df.drop(columns=['survey_id', 'title', 'block'])

    #
    # Split activity / pressure / state lists into separate rows, and reset index
    #

    for col in ['activity', 'pressure', 'state']:
        survey_df = survey_df.explode(column=col)
        survey_df = survey_df.reset_index(drop=True)

    #
    # Replace nan values with zeros and convert columns to integers
    #

    for column in ['measure', 'activity', 'pressure', 'state']:
        with warnings.catch_warnings(action='ignore'):
            survey_df[column] = survey_df[column].fillna(0)
        survey_df[column] = survey_df[column].astype(int)

    return survey_df

process_pressure_survey_data(file_name)

This method reads input from the excel file containing data about pressure contributions to states and the changes in state required to reach required thresholds of improvement.

Parameters:

Name Type Description Default
file_name str

path of survey excel file.

 required

Returns:

Name Type Description
pressure_contributions DataFrame
  • State: state id.
  • pressure: pressure id.
  • area_id: area id.
  • average: average contribution of pressure.
  • uncertainty: standard deviation of pressure contribution.
thresholds DataFrame
  • state: state id.
  • area_id: area id.
  • PR: reduction in state required to reach GES target.
  • 10: reduction in state required to reach 10 % improvement.
  • 25: reduction in state required to reach 25 % improvement.
  • 50: reduction in state required to reach 50 % improvement.
Source code in src/som_tools.py 
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
def process_pressure_survey_data(file_name: str) -> tuple[pd.DataFrame, pd.DataFrame]:
    """
    This method reads input from the excel file containing data about pressure contributions to states 
    and the changes in state required to reach required thresholds of improvement.

    Arguments:
        file_name (str): path of survey excel file.

    Returns:
        pressure_contributions (DataFrame):

            - State: state id.
            - pressure: pressure id.
            - area_id: area id.
            - average: average contribution of pressure.
            - uncertainty: standard deviation of pressure contribution.

        thresholds (DataFrame):

            - state: state id.
            - area_id: area id.
            - PR: reduction in state required to reach GES target.
            - 10: reduction in state required to reach 10 % improvement.
            - 25: reduction in state required to reach 25 % improvement.
            - 50: reduction in state required to reach 50 % improvement.
    """
    #
    # set parameter values
    #

    expert_number = 6   # max number of experts per question
    threshold_cols = ['PR', '10', '25', '50']   # target thresholds (PR=GES)

    #
    # read information sheet from input Excel file
    #

    data = pd.read_excel(io=file_name, sheet_name=None)
    sheet_names = list(data.keys())

    psq = data[sheet_names[0]]

    pressure_survey_data = {}
    for id in range(1, len(sheet_names)):
        pressure_survey_data[id] = data[sheet_names[id]]

    #
    # preprocess values
    #

    psq['area_id'] = [x.split(';') if type(x) == str else x for x in psq['area_id']]

    # add question id column
    psq['question_id'] = list(range(len(psq)))

    #
    # create new dataframe
    #

    survey_df = pd.DataFrame(columns=['survey_id', 'question_id', 'State', 'area_id', 'GES known', 'Weight'])

    # survey columns from which to take data
    cols = ['Expert']
    cols += [x + str(i+1) for x in ['P', 'S'] for i in range(expert_number)]    # up to expert_number different pressures related to state, and their significance
    cols += [x + y for x in ['MIN', 'MAX', 'ML'] for y in threshold_cols]     # required pressure reduction to reach GES (if known) or X % improvement in state

    start = 0    # keep track of where to access data in psq

    # for every survey sheet
    for survey_id in pressure_survey_data:

        # identify amount of experts in survey
        expert_ids = pressure_survey_data[survey_id]['Expert'].unique()
        # identify amount of questions in survey
        questions = np.sum(pressure_survey_data[survey_id]['Expert'] == expert_ids[0])

        # use number of questions to get state, area and GES known
        question_id = psq['question_id'].iloc[start:start+questions].reset_index(drop=True)
        state = psq['State'].iloc[start:start+questions].reset_index(drop=True)
        areas = psq['area_id'].iloc[start:start+questions].reset_index(drop=True)
        ges_known = psq['GES known'].iloc[start:start+questions].reset_index(drop=True)

        # find all expert weight columns and values
        expert_cols = [True if 'exp' in col.lower() else False for col in psq.columns]
        expert_weights = psq.loc[start:start+questions, expert_cols].reset_index(drop=True)
        expert_weights = expert_weights.fillna(1)

        survey_answers = 0
        for expert in expert_ids:

            # select expert answers
            data = pressure_survey_data[survey_id][cols].loc[pressure_survey_data[survey_id][cols]['Expert'] == expert].reset_index(drop=True)

            # verify that the amount of answers is correct
            if len(data) != questions: raise Exception('Not same amount of answers for each expert in survey sheet!')

            survey_answers += len(data)

            # set survey id, state, area and GES known for data
            data['survey_id'] = survey_id
            data['question_id'] = question_id
            data['State'] = state
            data['area_id'] = areas
            data['GES known'] = ges_known

            # set expert weights
            data['Weight'] = expert_weights['Exp' + str(int(expert))]

            # add data to final dataframe
            with warnings.catch_warnings(action='ignore'):
                survey_df = pd.concat([survey_df, data], ignore_index=True, sort=False)

        # verify that the correct number of answers was saved
        if survey_answers != len(expert_ids) * questions: raise Exception('Incorrect amount of answers found for survey!')

        # increase counter
        start += questions

    # create new dataframe for merged rows
    cols = ['survey_id', 'question_id', 'State', 'area_id', 'GES known']
    new_df = pd.DataFrame(columns=cols+['Pressures', 'Contribution']+threshold_cols)
    # remove empty elements from areas, and convert ids to integers
    survey_df['area_id'] = survey_df['area_id'].apply(lambda x: [int(area) for area in x if area != ''])
    # identify all unique questions
    questions = survey_df['question_id'].unique()
    # process each state
    for question in questions:
        # select current question rows
        data = survey_df.loc[survey_df['question_id'] == question].reset_index(drop=True)
        #
        # pressure contributions and uncertainties
        #
        # select pressures and significances, and find non nan values
        pressures = data[['P'+str(x+1) for x in range(expert_number)]].to_numpy().astype(float)
        significances = data[['S'+str(x+1) for x in range(expert_number)]].to_numpy().astype(float)
        mask = ~np.isnan(pressures)
        # weigh significances by amount of participating experts
        w = data[['Weight']].to_numpy().astype(float)
        significances = significances * w
        # go through each expert answer and calculate weights
        weights = {}
        for i, e in enumerate(pressures):
            s_tot = np.sum(significances[i][mask[i]])
            for p, s in zip(pressures[i][mask[i]], significances[i][mask[i]]):
                if int(p) not in weights:
                    weights[int(p)] = []
                weights[int(p)].append(s / s_tot)
        # using weights, calculate contributions and uncertainties
        average = {p: np.mean(weights[p]) for p in weights}
        stddev = {p: np.std(weights[p]) for p in weights}
        # create probability distributions
        minimum, maximum = {}, {}
        for p in average:
            if average[p] - stddev[p] > 0:
                if average[p] + stddev[p] > 1:
                    minimum[p] = 1.0 - stddev[p] * 2
                    maximum[p] = 1.0
                else:
                    minimum[p] = average[p] - stddev[p]
                    maximum[p] = average[p] + stddev[p]
            else:
                minimum[p] = 0.0
                maximum[p] = stddev[p] * 2
        average = {p: np.array([average[p] * 100]) for p in average}
        minimum = {p: np.array([minimum[p] * 100]) for p in minimum}
        maximum = {p: np.array([maximum[p] * 100]) for p in maximum}
        contribution = {p: get_prob_dist(average[p], minimum[p], maximum[p], np.ones(len(average[p]))) for p in average}
        # convert to lists
        pressures = list(average.keys())
        contribution = [contribution[p] for p in pressures]
        #
        # probability distributions for GES thresholds
        #
        reductions = {}
        for r in threshold_cols:
            # get min, max and ml data
            r_min = data['MIN'+r].to_numpy().astype(float)
            r_max = data['MAX'+r].to_numpy().astype(float)
            r_ml = data['ML'+r].to_numpy().astype(float)
            # get weighted cumulative probability distribution
            dist = get_prob_dist(r_ml, r_min, r_max, w.flatten())
            reductions[r] = dist
        #
        # merge processed data with dataframe
        #
        # create a new dataframe row and merge with new dataframe
        data = survey_df[cols].loc[survey_df['question_id'] == question].reset_index(drop=True).iloc[0]
        data = pd.DataFrame([data])
        # initialize new columns
        for c in ['Pressures', 'Contribution']+threshold_cols:
            data[c] = np.nan
            data[c] = data[c].astype(object)
        # change data type to allow for lists
        data.at[0, 'Pressures'] = pressures
        data.at[0, 'Contribution'] = contribution
        for r in threshold_cols:
            data.at[0, r] = reductions[r]
        with warnings.catch_warnings(action='ignore'):
            new_df = pd.concat([new_df, data], ignore_index=True, sort=False)
    #
    # split pressures into separate rows
    #
    new_df = new_df.assign(pressure=[list(zip(*row)) for row in zip(new_df['Pressures'], new_df['Contribution'])])
    new_df = new_df.explode('pressure')
    new_df = new_df.reset_index(drop=True)
    new_df = new_df.drop(columns=['Pressures', 'Contribution'])
    new_df[['pressure', 'contribution']] = pd.DataFrame(new_df['pressure'].tolist())
    #
    # remove rows with missing data (no pressure or no thresholds)
    #
    new_df = new_df.loc[new_df['pressure'].notna(), :]
    new_df = new_df[new_df[threshold_cols].notna().any(axis=1)]
    new_df = new_df.reset_index(drop=True)
    #
    # make sure pressure ids are integers, remove unnecessary columns
    #
    new_df['pressure'] = new_df['pressure'].astype(int)
    new_df = new_df.drop(columns=['survey_id', 'question_id', 'GES known'])
    #
    # split areas into separate rows
    #
    new_df = new_df.explode('area_id')
    new_df = new_df.reset_index(drop=True)
    #
    # split new_df into two dataframes, one for pressure contributions and one for thresholds
    #
    pressure_contributions = pd.DataFrame(new_df.loc[:, ['State', 'pressure', 'area_id', 'contribution']])
    thresholds = pd.DataFrame(new_df.loc[:, ['State', 'area_id'] + threshold_cols])
    #
    # rename capitalized columns
    #
    pressure_contributions = pressure_contributions.rename(columns={'State': 'state'})
    thresholds = thresholds.rename(columns={'State': 'state'})

    return pressure_contributions, thresholds

read_activity_contributions(file_name, sheet_name)

Reads input data of activity contributions to pressures in areas.

Parameters:

Name Type Description Default
file_name str

name of source excel file name.

 required
sheet_name str

name of excel sheet.

 required

Returns:

Name Type Description
act_to_press DataFrame

dataframe containing mappings between activities and pressures.
Source code in src/som_tools.py 
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
def read_activity_contributions(file_name: str, sheet_name: str) -> pd.DataFrame:
    """
    Reads input data of activity contributions to pressures in areas. 

    Arguments:
        file_name (str): name of source excel file name.
        sheet_name (str): name of excel sheet.

    Returns:
        act_to_press (DataFrame): dataframe containing mappings between activities and pressures.
    """
    act_to_press = pd.read_excel(file_name, sheet_name=sheet_name)

    # read all most likely, min and max column values into lists in new columns
    for col, regex_str in zip(['expected', 'minimum', 'maximum'], ['ML[1-6]', 'Min[1-6]', 'Max[1-6]']):
        act_to_press[col] = act_to_press.filter(regex=regex_str).values.tolist()

    # remove all most likely, min and max columns
    for regex_str in ['ML[1-6]', 'Min[1-6]', 'Max[1-6]']:
        act_to_press.drop(act_to_press.filter(regex=regex_str).columns, axis=1, inplace=True)

    # separate values grouped together in sheet on the same row with ';' into separate rows
    for category in ['Activity', 'Pressure', 'area_id']:
        act_to_press[category] = [list(filter(None, x.split(';'))) if type(x) == str else x for x in act_to_press[category]]
        act_to_press = act_to_press.explode(category)
        act_to_press[category] = act_to_press[category].astype(int)
    act_to_press = act_to_press.reset_index(drop=True)

    # calculate probability distributions
    act_to_press['contribution'] = pd.Series([np.nan] * len(act_to_press), dtype='object')
    for num in act_to_press.index:
        # convert expert answers to array
        expected = np.array(list(act_to_press.loc[num, ['expected']])).flatten()
        lower = np.array(list(act_to_press.loc[num, ['minimum']])).flatten()
        upper = np.array(list(act_to_press.loc[num, ['maximum']])).flatten()
        weights = np.full(len(expected), 1)
        # if boundaries are unknown, set to same as expected
        lower[np.isnan(lower)] = expected[np.isnan(lower)]
        upper[np.isnan(upper)] = expected[np.isnan(upper)]
        # get probability distribution
        act_to_press.at[num, 'contribution'] = get_prob_dist(expected, lower, upper, weights)

    act_to_press = act_to_press.drop(columns=['expected', 'minimum', 'maximum'])

    # rename columns
    act_to_press = act_to_press.rename(columns={'Activity': 'activity', 'Pressure': 'pressure'})

    return act_to_press

read_cases(file_name, sheet_name)

Reads input data for cases. Each row represents one case.

In columns of 'ActMeas' sheet ('activities', 'pressure' and 'state') the value 0 == 'all relevant'.

Parameters:

Name Type Description Default
file_name str

name of source excel file name.

 required
sheet_name str

name of excel sheet.

 required

Returns:

Name Type Description
cases DataFrame

case data.
Source code in src/som_tools.py 
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
def read_cases(file_name: str, sheet_name: str) -> pd.DataFrame:
    """
    Reads input data for cases. Each row represents one case. 

    In columns of 'ActMeas' sheet ('activities', 'pressure' and 'state') the value 0 == 'all relevant'.

    Arguments:
        file_name (str): name of source excel file name.
        sheet_name (str): name of excel sheet.

    Returns:
        cases (DataFrame): case data.
    """
    cases = pd.read_excel(io=file_name, sheet_name=sheet_name)

    assert len(cases[cases.duplicated(['ID'])]) == 0

    for col in ['activity', 'pressure', 'state', 'area_id']:
        # separate ids grouped together in sheet on the same row with ';' into separate rows
        cases[col] = [list(filter(None, x.split(';'))) if type(x) == str else x for x in cases[col]]
        cases = cases.explode(col)
        # change types of split values from str to int
        cases[col] = cases[col].astype(int)

    for col in ['coverage', 'implementation']:
        cases[col] = cases[col].astype(float)

    cases = cases.reset_index(drop=True)

    return cases

read_development_scenarios(file_name, sheet_name)

Reads input data of activity development scenarios.

Parameters:

Name Type Description Default
file_name str

name of source excel file name.

 required
sheet_name str

name of sheet in excel file.

 required

Returns:

Name Type Description
development_scenarios DataFrame

dataframe containing activity development scenarios.
Source code in src/som_tools.py 
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
def read_development_scenarios(file_name: str, sheet_name: str) -> pd.DataFrame:
    """
    Reads input data of activity development scenarios. 

    Arguments:
        file_name (str): name of source excel file name.
        sheet_name (str): name of sheet in excel file.

    Returns:
        development_scenarios (DataFrame): dataframe containing activity development scenarios.
    """
    development_scenarios = pd.read_excel(file_name, sheet_name=sheet_name)

    # replace nan values with 0, assuming that no value means no change
    for category in ['BAU', 'ChangeMin', 'ChangeML', 'ChangeMax']:
        development_scenarios.loc[np.isnan(development_scenarios[category]), category] = 0
        development_scenarios[category] = development_scenarios[category].astype(float)

    development_scenarios['Activity'] = development_scenarios['Activity'].astype(int)

    # change values from percentual change to multiplier type by adding 1
    for category in ['BAU', 'ChangeMin', 'ChangeML', 'ChangeMax']:
        development_scenarios[category] = development_scenarios[category] + 1

    # rename column
    development_scenarios = development_scenarios.rename(columns={'Activity': 'activity'})

    return development_scenarios

read_ids(file_name, id_sheets)

Reads in model component ids and descriptions from general input files.

Parameters:

Name Type Description Default
file_name str

source excel file name containing measure, activity, pressure and state id sheets.

 required
id_sheets dict

should have structure {'measure': sheet_name, 'activity': sheet_name, ...}.

 required

Returns:

Name Type Description
object_data dict

dictionary containing measure, activity, pressure and state ids and descriptions in separate dataframes.
Source code in src/som_tools.py 
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
def read_ids(file_name: str, id_sheets: dict) -> dict[str, dict]:
    """
    Reads in model component ids and descriptions from general input files.

    Arguments:
        file_name (str): source excel file name containing measure, activity, pressure and state id sheets.
        id_sheets (dict): should have structure {'measure': sheet_name, 'activity': sheet_name, ...}.

    Returns:
        object_data (dict): dictionary containing measure, activity, pressure and state ids and descriptions in separate dataframes.
    """
    # create dicts for each category
    object_data = {}
    for category in id_sheets:
        # read excel sheet into dataframe
        df = pd.read_excel(io=file_name, sheet_name=id_sheets[category])
        # remove non-necessary columns
        df.drop(columns=[col for col in df.columns if col not in ['ID', category]])
        # remove rows where id is nan or empty string
        df = df.dropna(subset=['ID'])
        df = df[df['ID'] != '']
        # convert id column to integer (if not already)
        df['ID'] = df['ID'].astype(int)
        object_data[category] = df

    return object_data

read_overlaps(file_name, sheet_name)

Reads input data of measure-measure interactions.

Parameters:

Name Type Description Default
file_name str

name of source excel file name.

 required
sheet_name str

name of sheet in excel file.

 required

Returns:

Name Type Description
overlaps DataFrame

dataframe containing overlaps between individual measures.
Source code in src/som_tools.py 
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
def read_overlaps(file_name: str, sheet_name: str) -> pd.DataFrame:
    """
    Reads input data of measure-measure interactions. 

    Arguments:
        file_name (str): name of source excel file name.
        sheet_name (str): name of sheet in excel file.

    Returns:
        overlaps (DataFrame): dataframe containing overlaps between individual measures.
    """
    overlaps = pd.read_excel(file_name, sheet_name=sheet_name)

    # replace nan values in ID columns with 0 and make sure they are integers
    for category in ['Overlap', 'Pressure', 'Activity', 'Overlapping', 'Overlapped']:
        overlaps.loc[np.isnan(overlaps[category]), category] = 0
        overlaps[category] = overlaps[category].astype(int)

    overlaps = overlaps.rename(columns={'Overlap': 'overlap', 
                                'Pressure': 'pressure', 
                                'Activity': 'activity', 
                                'Overlapping': 'overlapping', 
                                'Overlapped': 'overlapped', 
                                'Multiplier': 'multiplier'})

    return overlaps

read_subpressures(file_name, sheet_name)

Reads input data of subpressures links to state pressures.

Parameters:

Name Type Description Default
file_name str

name of source excel file name.

 required
sheet_name str

name of sheet in excel file.

 required

Returns:

Name Type Description
subpressures DataFrame

dataframe containing subpressure links.
Source code in src/som_tools.py 
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
def read_subpressures(file_name: str, sheet_name: str) -> pd.DataFrame:
    """
    Reads input data of subpressures links to state pressures.

    Arguments:
        file_name (str): name of source excel file name.
        sheet_name (str): name of sheet in excel file.

    Returns:
        subpressures (DataFrame): dataframe containing subpressure links.
    """
    subpressures = pd.read_excel(file_name, sheet_name=sheet_name)

    for col in ['Reduced pressure', 'State pressure', 'State']:
        # separate ids grouped together in sheet on the same row with ';' into separate rows
        subpressures[col] = [list(filter(None, x.split(';'))) if type(x) == str else x for x in subpressures[col]]
        subpressures = subpressures.explode(col)
        # change types of split values from str to int
        subpressures[col] = subpressures[col].astype(int)

    for col in subpressures.columns:
        if col not in ['Reduced pressure', 'State pressure', 'State', 'Equivalence']:
            subpressures = subpressures.drop(columns=[col])

    subpressures = subpressures.reset_index(drop=True)

    def assign_multiplier(equivalence):
        if equivalence <= 1:
            return equivalence
        elif equivalence == 2:
            return 0
        elif equivalence == 3:
            return 0
        else:
            return 0

    subpressures['Multiplier'] = subpressures['Equivalence'].apply(assign_multiplier)

    # rename columns
    subpressures = subpressures.rename(columns={'Reduced pressure': 'reduced pressure', 
                                        'State pressure': 'state pressure', 
                                        'Equivalence': 'equivalence', 
                                        'State': 'state', 
                                        'Multiplier': 'multiplier'})

    return subpressures