Google_Demo_Attribution(MAM).ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"Post_blog_Attribution(MAM).ipynb","provenance":[{"file_id":"1AHWrqMC30tL-7lSYi2XrNBvdCy6wdfoO","timestamp":1590962763503}],"collapsed_sections":["3-dC9Tjg7Y3F","5jph_mBG8m6P"],"toc_visible":true,"machine_shape":"hm"},"kernelspec":{"name":"python3","display_name":"Python 3"}},"cells":[{"cell_type":"markdown","metadata":{"id":"5jph_mBG8m6P","colab_type":"text"},"source":["## Demostração Blog DP6"]},{"cell_type":"markdown","metadata":{"id":"QMi0dcAd0ldr","colab_type":"text"},"source":["Importando pacotes necessários:"]},{"cell_type":"code","metadata":{"id":"cQJ60vzE0Zpm","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":313},"executionInfo":{"status":"ok","timestamp":1593454212947,"user_tz":180,"elapsed":6678,"user":{"displayName":"André Tocci Russi","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjUOmZJ5GKqIq-1s5NcPvDx4AUU0mOWUZw1W62r=s64","userId":"17444405581757145872"}},"outputId":"41b3e91d-a51a-4e07-dbbd-816ac65d8a0d"},"source":[" import pandas as pd\n","import numpy as np\n","import matplotlib.pyplot as plt\n","import seaborn as sns\n","\n","!pip install marketing_attribution_models\n","from marketing_attribution_models import MAM"],"execution_count":null,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.6/dist-packages/statsmodels/tools/_testing.py:19: FutureWarning: pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.\n","  import pandas.util.testing as tm\n"],"name":"stderr"},{"output_type":"stream","text":["Collecting marketing_attribution_models\n","  Downloading https://files.pythonhosted.org/packages/f9/81/b76f9f7aa2f1ceb2f5380012840363d37ff3cee72c0a497d3ec253430d59/marketing_attribution_models-1.0.7-py3-none-any.whl\n","Requirement already satisfied: seaborn in /usr/local/lib/python3.6/dist-packages (from marketing_attribution_models) (0.10.1)\n","Requirement already satisfied: numpy in /usr/local/lib/python3.6/dist-packages (from marketing_attribution_models) (1.18.5)\n","Requirement already satisfied: matplotlib in /usr/local/lib/python3.6/dist-packages (from marketing_attribution_models) (3.2.2)\n","Requirement already satisfied: pandas in /usr/local/lib/python3.6/dist-packages (from marketing_attribution_models) (1.0.5)\n","Requirement already satisfied: scipy>=1.0.1 in /usr/local/lib/python3.6/dist-packages (from seaborn->marketing_attribution_models) (1.4.1)\n","Requirement already satisfied: python-dateutil>=2.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib->marketing_attribution_models) (2.8.1)\n","Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.6/dist-packages (from matplotlib->marketing_attribution_models) (0.10.0)\n","Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib->marketing_attribution_models) (1.2.0)\n","Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib->marketing_attribution_models) (2.4.7)\n","Requirement already satisfied: pytz>=2017.2 in /usr/local/lib/python3.6/dist-packages (from pandas->marketing_attribution_models) (2018.9)\n","Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.6/dist-packages (from python-dateutil>=2.1->matplotlib->marketing_attribution_models) (1.12.0)\n","Installing collected packages: marketing-attribution-models\n","Successfully installed marketing-attribution-models-1.0.7\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"YYXBTN-fDWGc","colab_type":"code","cellView":"form","colab":{"base_uri":"https://localhost:8080/","height":118},"executionInfo":{"status":"ok","timestamp":1593454272967,"user_tz":180,"elapsed":24694,"user":{"displayName":"André Tocci Russi","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjUOmZJ5GKqIq-1s5NcPvDx4AUU0mOWUZw1W62r=s64","userId":"17444405581757145872"}},"outputId":"7c9db162-8315-47a1-e2d1-9d6987e376e6"},"source":["#@title Importando e Tratando as Bases\n","import glob\n","from google.colab import drive\n","drive.mount('/content/drive/')\n","\n","filepath = '/content/drive/Shared drives/Data Science/Ultron/Recipes/DP6tribution/'\n","filename = '*'\n","all_files = glob.glob(filepath + filename)\n","pd.set_option('display.float_format', lambda x: '%.3f' % x)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3aietf%3awg%3aoauth%3a2.0%3aoob&response_type=code&scope=email%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdocs.test%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive.photos.readonly%20https%3a%2f%2fwww.googleapis.com%2fauth%2fpeopleapi.readonly\n","\n","Enter your authorization code:\n","··········\n","Mounted at /content/drive/\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"3wFNrzbniHdx","colab_type":"text"},"source":["Para está demonstração, optamos por demonstrar utilizando uma base de dados bastante conhecida pelos profissionais do Marketing: Google Merchandise Store. Está base foi extraída da conta de demonstração do Google Analytics, que pode ser acessada por qualquer usuário do Google."]},{"cell_type":"code","metadata":{"id":"7KF3LATxhecU","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":191},"executionInfo":{"status":"ok","timestamp":1593454275134,"user_tz":180,"elapsed":2159,"user":{"displayName":"André Tocci Russi","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjUOmZJ5GKqIq-1s5NcPvDx4AUU0mOWUZw1W62r=s64","userId":"17444405581757145872"}},"outputId":"07d236d4-8e61-4187-faf4-334d4dc0de9a"},"source":["#Lendo os dados do Google Merchandise Store, mas selecionando apenas as colunas que serão utilizadas para gerar os modelos\n","df = pd.read_csv(all_files[0], \n","                 usecols=['channelGrouping', \n","                          'date', \n","                          'fullVisitorId', \n","                          'totals_transactions', \n","                          'totals_transactionRevenue'],\n","                 dtype = {'channelGrouping':str, \n","                          'fullVisitorId':str, \n","                          'totals_transactions' : np.float64, \n","                          'totals_transactionRevenue' : np.float64} )\n","\n","#Analizando o DF retornado\n","df.head()"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>channelGrouping</th>\n","      <th>date</th>\n","      <th>fullVisitorId</th>\n","      <th>totals_transactions</th>\n","      <th>totals_transactionRevenue</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>Organic Search</td>\n","      <td>20171016</td>\n","      <td>579247845938461901</td>\n","      <td>nan</td>\n","      <td>nan</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>Organic Search</td>\n","      <td>20171016</td>\n","      <td>875141984250977091</td>\n","      <td>nan</td>\n","      <td>nan</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>Organic Search</td>\n","      <td>20171016</td>\n","      <td>953429090981710815</td>\n","      <td>nan</td>\n","      <td>nan</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>Organic Search</td>\n","      <td>20171016</td>\n","      <td>2365540140634612791</td>\n","      <td>nan</td>\n","      <td>nan</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>Organic Search</td>\n","      <td>20171016</td>\n","      <td>8585866264807101623</td>\n","      <td>nan</td>\n","      <td>nan</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["  channelGrouping      date  ... totals_transactions  totals_transactionRevenue\n","0  Organic Search  20171016  ...                 nan                        nan\n","1  Organic Search  20171016  ...                 nan                        nan\n","2  Organic Search  20171016  ...                 nan                        nan\n","3  Organic Search  20171016  ...                 nan                        nan\n","4  Organic Search  20171016  ...                 nan                        nan\n","\n","[5 rows x 5 columns]"]},"metadata":{"tags":[]},"execution_count":4}]},{"cell_type":"markdown","metadata":{"id":"vSLwAAiPAZuI","colab_type":"text"},"source":["Mas antes de aplicarmos o modelo utilizando a MAM, precisamos realizar alguns ajustes na base de dados. \n","\n","E analisando o .head() e o .info(), conseguimos observar que:\n","\n","- A coluna de data não está no formato datetime;\n","- As colunas totals_transactions e totals_transactionRevenue possuem valores vazios;\n","- fullVisitorId foi lido como objeto, temos que tomar cuidado para não ler essa informação como número, 001 pode ser um ID diferente de1; \n","- A base não possui nenhum agrupamento de jornada;"]},{"cell_type":"code","metadata":{"id":"HizeMfsgzOML","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":212},"executionInfo":{"status":"ok","timestamp":1590963551094,"user_tz":180,"elapsed":641,"user":{"displayName":"André Tocci Russi","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjUOmZJ5GKqIq-1s5NcPvDx4AUU0mOWUZw1W62r=s64","userId":"17444405581757145872"}},"outputId":"4692e258-7957-4c2d-9d30-06736f53159d"},"source":["#Selecionando colunas necessárias\n","df.info()"],"execution_count":null,"outputs":[{"output_type":"stream","text":["<class 'pandas.core.frame.DataFrame'>\n","RangeIndex: 73948 entries, 0 to 73947\n","Data columns (total 5 columns):\n"," #   Column                     Non-Null Count  Dtype  \n","---  ------                     --------------  -----  \n"," 0   channelGrouping            73948 non-null  object \n"," 1   date                       73948 non-null  int64  \n"," 2   fullVisitorId              73948 non-null  object \n"," 3   totals_transactions        818 non-null    float64\n"," 4   totals_transactionRevenue  817 non-null    float64\n","dtypes: float64(2), int64(1), object(2)\n","memory usage: 2.8+ MB\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"eM8oC82IlEfz","colab_type":"text"},"source":["Assim, tratamos a coluna de data para o formato correto, tratamos os valores nulos e criamos uma coluna indicando se houve ou não uma conversão naquela sessão."]},{"cell_type":"code","metadata":{"id":"IgowOnl20_-s","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":191},"executionInfo":{"status":"ok","timestamp":1593454309372,"user_tz":180,"elapsed":557,"user":{"displayName":"André Tocci Russi","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjUOmZJ5GKqIq-1s5NcPvDx4AUU0mOWUZw1W62r=s64","userId":"17444405581757145872"}},"outputId":"10a3de01-c6ab-4c8c-f913-b275119a9c06"},"source":["df['date_tratada']  = pd.to_datetime( df['date'], format='%Y%m%d' )\n","df['totals_transactions'].fillna(0, inplace=True)\n","df['totals_transactionRevenue'].fillna(0, inplace=True)\n","df['has_transaction'] = df.totals_transactions.apply(lambda x: True if x > 0 else False)\n","\n","df.head()"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>channelGrouping</th>\n","      <th>date</th>\n","      <th>fullVisitorId</th>\n","      <th>totals_transactions</th>\n","      <th>totals_transactionRevenue</th>\n","      <th>date_tratada</th>\n","      <th>has_transaction</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>Organic Search</td>\n","      <td>20171016</td>\n","      <td>579247845938461901</td>\n","      <td>0.000</td>\n","      <td>0.000</td>\n","      <td>2017-10-16</td>\n","      <td>False</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>Organic Search</td>\n","      <td>20171016</td>\n","      <td>875141984250977091</td>\n","      <td>0.000</td>\n","      <td>0.000</td>\n","      <td>2017-10-16</td>\n","      <td>False</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>Organic Search</td>\n","      <td>20171016</td>\n","      <td>953429090981710815</td>\n","      <td>0.000</td>\n","      <td>0.000</td>\n","      <td>2017-10-16</td>\n","      <td>False</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>Organic Search</td>\n","      <td>20171016</td>\n","      <td>2365540140634612791</td>\n","      <td>0.000</td>\n","      <td>0.000</td>\n","      <td>2017-10-16</td>\n","      <td>False</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>Organic Search</td>\n","      <td>20171016</td>\n","      <td>8585866264807101623</td>\n","      <td>0.000</td>\n","      <td>0.000</td>\n","      <td>2017-10-16</td>\n","      <td>False</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["  channelGrouping      date  ... date_tratada  has_transaction\n","0  Organic Search  20171016  ...   2017-10-16            False\n","1  Organic Search  20171016  ...   2017-10-16            False\n","2  Organic Search  20171016  ...   2017-10-16            False\n","3  Organic Search  20171016  ...   2017-10-16            False\n","4  Organic Search  20171016  ...   2017-10-16            False\n","\n","[5 rows x 7 columns]"]},"metadata":{"tags":[]},"execution_count":5}]},{"cell_type":"markdown","metadata":{"id":"uQZP8skv2LZV","colab_type":"text"},"source":["Agora que tratamos nossa base, podemos criar o nosso objeto MAM, como nossa base está na granularidade de sessão por usuário e não possuímos um ID de agrupamento de jornada, é importante que passamos como True os parametros group_channels e create_journey_id_based_on_conversion."]},{"cell_type":"code","metadata":{"id":"WXvVsowi2Qks","colab_type":"code","colab":{}},"source":["DP_tribution = MAM(df, \n","                  channels_colname='channelGrouping',\n","                  group_channels=True, \n","                  group_channels_by_id_list=['fullVisitorId'], \n","                  group_timestamp_colname='date_tratada',\n","                  journey_with_conv_colname='has_transaction',\n","                  create_journey_id_based_on_conversion = True,\n","                  conversion_value='totals_transactionRevenue',)"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"Jad_s47WnmB-","colab_type":"text"},"source":["Acessando o atributo .DataFrame, conseguimos observar que houve uma mudança significativa em nossa base de dados. Observamos que a granularidade da tabela foi alterada de sessão para jornada e que há uma coluna nova de id da jornada."]},{"cell_type":"code","metadata":{"id":"SsvxcV0EAiYf","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":191},"executionInfo":{"status":"ok","timestamp":1590963765109,"user_tz":180,"elapsed":616,"user":{"displayName":"André Tocci Russi","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjUOmZJ5GKqIq-1s5NcPvDx4AUU0mOWUZw1W62r=s64","userId":"17444405581757145872"}},"outputId":"8df93d34-95f2-44bd-8914-ff11de8b66f3"},"source":["DP_tribution.DataFrame.sort_values('conversion_value', ascending=False).head()"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>journey_id</th>\n","      <th>channels_agg</th>\n","      <th>converted_agg</th>\n","      <th>conversion_value</th>\n","      <th>time_till_conv_agg</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>12649</th>\n","      <td>id:2969957690756250633_J:0</td>\n","      <td>Organic Search &gt; Organic Search &gt; Organic Sear...</td>\n","      <td>True</td>\n","      <td>4198500000.000</td>\n","      <td>72.0 &gt; 48.0 &gt; 24.0 &gt; 0.0</td>\n","    </tr>\n","    <tr>\n","      <th>48276</th>\n","      <td>id:8539433516671325668_J:0</td>\n","      <td>Direct &gt; Direct &gt; Direct &gt; Direct &gt; Direct &gt; D...</td>\n","      <td>True</td>\n","      <td>2870000000.000</td>\n","      <td>48.0 &gt; 48.0 &gt; 24.0 &gt; 24.0 &gt; 0.0 &gt; 0.0 &gt; 0.0</td>\n","    </tr>\n","    <tr>\n","      <th>12230</th>\n","      <td>id:2904329822175970539_J:4</td>\n","      <td>Referral &gt; Referral &gt; Referral</td>\n","      <td>True</td>\n","      <td>2578500000.000</td>\n","      <td>600.0 &gt; 288.0 &gt; 0.0</td>\n","    </tr>\n","    <tr>\n","      <th>41360</th>\n","      <td>id:7445235885559107095_J:3</td>\n","      <td>Organic Search &gt; Organic Search &gt; Organic Sear...</td>\n","      <td>True</td>\n","      <td>2490000000.000</td>\n","      <td>768.0 &gt; 408.0 &gt; 288.0 &gt; 264.0 &gt; 264.0 &gt; 96.0 &gt;...</td>\n","    </tr>\n","    <tr>\n","      <th>26992</th>\n","      <td>id:5183829074529712735_J:0</td>\n","      <td>Direct &gt; Direct &gt; Direct</td>\n","      <td>True</td>\n","      <td>1919200000.000</td>\n","      <td>0.0 &gt; 0.0 &gt; 0.0</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                       journey_id  ...                                 time_till_conv_agg\n","12649  id:2969957690756250633_J:0  ...                           72.0 > 48.0 > 24.0 > 0.0\n","48276  id:8539433516671325668_J:0  ...        48.0 > 48.0 > 24.0 > 24.0 > 0.0 > 0.0 > 0.0\n","12230  id:2904329822175970539_J:4  ...                                600.0 > 288.0 > 0.0\n","41360  id:7445235885559107095_J:3  ...  768.0 > 408.0 > 288.0 > 264.0 > 264.0 > 96.0 >...\n","26992  id:5183829074529712735_J:0  ...                                    0.0 > 0.0 > 0.0\n","\n","[5 rows x 5 columns]"]},"metadata":{"tags":[]},"execution_count":10}]},{"cell_type":"markdown","metadata":{"id":"CwEldUXi1azw","colab_type":"text"},"source":["### Aplicando os Modelos"]},{"cell_type":"markdown","metadata":{"id":"pWJaeXGLoOLT","colab_type":"text"},"source":["Para rodar o modelo do Shapley Value, basta aplicar o método .attribution_shapley().\n","\n","O parâmetro **size limita quantidade de canais únicos na jornada**, por **padrão** é definido como os **4 últimos**. Isso ocorre pois o número de iterações aumenta exponencialmente com o número de canais. Da ordem de 2N, sendo N o número de canais.   \n","\n","A metodologia do cálculo das contribuições marginais pode variar através do **parâmetro order**, que por padrão calcula a contribuição da **combinação dos canais independende da ordem em que aparecem** nas diferentes jornadas.\n","\n","Por fim, parâmetro na qual o Shapley Value será calculado pode ser alterado em **values_col**, que por padrão utiliza a **taxa de conversão** que é uma forma de **considerarmos as não conversões no cálculo do modelo**. Contudo, também podemos considerar no cálculo o total de conversões ou o valor gerados pelas conversões, como demostrado abaixo. "]},{"cell_type":"code","metadata":{"id":"MIkMyKL0oOV7","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":472},"executionInfo":{"status":"ok","timestamp":1593454681197,"user_tz":180,"elapsed":817,"user":{"displayName":"André Tocci Russi","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjUOmZJ5GKqIq-1s5NcPvDx4AUU0mOWUZw1W62r=s64","userId":"17444405581757145872"}},"outputId":"f4919c27-6e83-4d4d-a0cd-3a0299487e21"},"source":["shapley_results = DP_tribution.attribution_shapley()"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(                               combinations  ...    attribution_shapley_size4_conv_rate_algorithmic\n"," 16                                   Direct  ...                                    [29708520000.0]\n"," 17                         Direct > Display  ...           [243368413.16711098, 199221586.83288905]\n"," 18        Direct > Display > Organic Search  ...  [15685297.311649714, 14457030.40197919, 383767...\n"," 21                  Direct > Organic Search  ...                [1086350193.20019, 725719806.79981]\n"," 24       Direct > Organic Search > Referral  ...  [334937900.0109693, 179302753.22538412, 896279...\n"," 28                        Direct > Referral  ...           [1653725008.2510033, 4242104991.7489963]\n"," 31                                  Display  ...                                      [238200000.0]\n"," 32                 Display > Organic Search  ...              [57058012.82148681, 76831987.1785132]\n"," 35      Display > Organic Search > Referral  ...  [7391834.050253105, 2623250.57365777, 16384915...\n"," 38                       Display > Referral  ...             [33501108.56390939, 80708891.43609062]\n"," 39                           Organic Search  ...                                    [25418020000.0]\n"," 40             Organic Search > Paid Search  ...            [116152975.34855956, 165957024.6514404]\n"," 41  Organic Search > Paid Search > Referral  ...  [25725024.54588916, 43524361.64828748, 7424061...\n"," 44                Organic Search > Referral  ...              [674082087.187786, 2860957912.812214]\n"," 47                              Paid Search  ...                                     [2282920000.0]\n"," 48                   Paid Search > Referral  ...            [225600785.96289113, 463269214.0371089]\n"," 50                                 Referral  ...                                    [40676950000.0]\n"," 52                                   Social  ...                                       [61080000.0]\n"," \n"," [18 rows x 6 columns], channels\n"," Direct           33042586811.941\n"," Display            549829572.671\n"," Organic Search   27222295557.146\n"," Paid Search       2718002172.263\n"," Referral         49310895885.980\n"," Social              61080000.000\n"," Name: value, dtype: float64)"]},"metadata":{"tags":[]},"execution_count":7}]},{"cell_type":"markdown","metadata":{"id":"k8W3qV1yuVGJ","colab_type":"text"},"source":["O método retorna uma tupla contendo :\n","- Os resultados agrupados por jornadas únicas;"]},{"cell_type":"code","metadata":{"id":"kac7Odr7uUbD","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":565},"executionInfo":{"status":"ok","timestamp":1590963772639,"user_tz":180,"elapsed":641,"user":{"displayName":"André Tocci Russi","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjUOmZJ5GKqIq-1s5NcPvDx4AUU0mOWUZw1W62r=s64","userId":"17444405581757145872"}},"outputId":"000d13c7-0f62-4aa9-c04d-f726beb67df1"},"source":["shapley_results[0]"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>combinations</th>\n","      <th>conversions</th>\n","      <th>total_sequences</th>\n","      <th>conversion_value</th>\n","      <th>conv_rate</th>\n","      <th>attribution_shapley_size4_conv_rate_algorithmic</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>16</th>\n","      <td>Direct</td>\n","      <td>157</td>\n","      <td>8647</td>\n","      <td>29708520000.000</td>\n","      <td>0.018</td>\n","      <td>[29708520000.0]</td>\n","    </tr>\n","    <tr>\n","      <th>17</th>\n","      <td>Direct &gt; Display</td>\n","      <td>2</td>\n","      <td>13</td>\n","      <td>442590000.000</td>\n","      <td>0.154</td>\n","      <td>[243368413.16711098, 199221586.83288905]</td>\n","    </tr>\n","    <tr>\n","      <th>18</th>\n","      <td>Direct &gt; Display &gt; Organic Search</td>\n","      <td>1</td>\n","      <td>6</td>\n","      <td>33980000.000</td>\n","      <td>0.167</td>\n","      <td>[15685297.311649714, 14457030.40197919, 383767...</td>\n","    </tr>\n","    <tr>\n","      <th>21</th>\n","      <td>Direct &gt; Organic Search</td>\n","      <td>9</td>\n","      <td>210</td>\n","      <td>1812070000.000</td>\n","      <td>0.043</td>\n","      <td>[1086350193.20019, 725719806.79981]</td>\n","    </tr>\n","    <tr>\n","      <th>24</th>\n","      <td>Direct &gt; Organic Search &gt; Referral</td>\n","      <td>3</td>\n","      <td>28</td>\n","      <td>1410520000.000</td>\n","      <td>0.107</td>\n","      <td>[334937900.0109693, 179302753.22538412, 896279...</td>\n","    </tr>\n","    <tr>\n","      <th>28</th>\n","      <td>Direct &gt; Referral</td>\n","      <td>22</td>\n","      <td>221</td>\n","      <td>5895830000.000</td>\n","      <td>0.100</td>\n","      <td>[1653725008.2510033, 4242104991.7489963]</td>\n","    </tr>\n","    <tr>\n","      <th>31</th>\n","      <td>Display</td>\n","      <td>4</td>\n","      <td>1423</td>\n","      <td>238200000.000</td>\n","      <td>0.003</td>\n","      <td>[238200000.0]</td>\n","    </tr>\n","    <tr>\n","      <th>32</th>\n","      <td>Display &gt; Organic Search</td>\n","      <td>3</td>\n","      <td>65</td>\n","      <td>133890000.000</td>\n","      <td>0.046</td>\n","      <td>[57058012.82148681, 76831987.1785132]</td>\n","    </tr>\n","    <tr>\n","      <th>35</th>\n","      <td>Display &gt; Organic Search &gt; Referral</td>\n","      <td>1</td>\n","      <td>7</td>\n","      <td>26400000.000</td>\n","      <td>0.143</td>\n","      <td>[7391834.050253105, 2623250.57365777, 16384915...</td>\n","    </tr>\n","    <tr>\n","      <th>38</th>\n","      <td>Display &gt; Referral</td>\n","      <td>3</td>\n","      <td>21</td>\n","      <td>114210000.000</td>\n","      <td>0.143</td>\n","      <td>[33501108.56390939, 80708891.43609062]</td>\n","    </tr>\n","    <tr>\n","      <th>39</th>\n","      <td>Organic Search</td>\n","      <td>240</td>\n","      <td>24929</td>\n","      <td>25418020000.000</td>\n","      <td>0.010</td>\n","      <td>[25418020000.0]</td>\n","    </tr>\n","    <tr>\n","      <th>40</th>\n","      <td>Organic Search &gt; Paid Search</td>\n","      <td>9</td>\n","      <td>255</td>\n","      <td>282110000.000</td>\n","      <td>0.035</td>\n","      <td>[116152975.34855956, 165957024.6514404]</td>\n","    </tr>\n","    <tr>\n","      <th>41</th>\n","      <td>Organic Search &gt; Paid Search &gt; Referral</td>\n","      <td>2</td>\n","      <td>9</td>\n","      <td>143490000.000</td>\n","      <td>0.222</td>\n","      <td>[25725024.54588916, 43524361.64828748, 7424061...</td>\n","    </tr>\n","    <tr>\n","      <th>44</th>\n","      <td>Organic Search &gt; Referral</td>\n","      <td>32</td>\n","      <td>379</td>\n","      <td>3535040000.000</td>\n","      <td>0.084</td>\n","      <td>[674082087.187786, 2860957912.812214]</td>\n","    </tr>\n","    <tr>\n","      <th>47</th>\n","      <td>Paid Search</td>\n","      <td>17</td>\n","      <td>1072</td>\n","      <td>2282920000.000</td>\n","      <td>0.016</td>\n","      <td>[2282920000.0]</td>\n","    </tr>\n","    <tr>\n","      <th>48</th>\n","      <td>Paid Search &gt; Referral</td>\n","      <td>2</td>\n","      <td>15</td>\n","      <td>688870000.000</td>\n","      <td>0.133</td>\n","      <td>[225600785.96289113, 463269214.0371089]</td>\n","    </tr>\n","    <tr>\n","      <th>50</th>\n","      <td>Referral</td>\n","      <td>308</td>\n","      <td>4979</td>\n","      <td>40676950000.000</td>\n","      <td>0.062</td>\n","      <td>[40676950000.0]</td>\n","    </tr>\n","    <tr>\n","      <th>52</th>\n","      <td>Social</td>\n","      <td>3</td>\n","      <td>14025</td>\n","      <td>61080000.000</td>\n","      <td>0.000</td>\n","      <td>[61080000.0]</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["                               combinations  ...    attribution_shapley_size4_conv_rate_algorithmic\n","16                                   Direct  ...                                    [29708520000.0]\n","17                         Direct > Display  ...           [243368413.16711098, 199221586.83288905]\n","18        Direct > Display > Organic Search  ...  [15685297.311649714, 14457030.40197919, 383767...\n","21                  Direct > Organic Search  ...                [1086350193.20019, 725719806.79981]\n","24       Direct > Organic Search > Referral  ...  [334937900.0109693, 179302753.22538412, 896279...\n","28                        Direct > Referral  ...           [1653725008.2510033, 4242104991.7489963]\n","31                                  Display  ...                                      [238200000.0]\n","32                 Display > Organic Search  ...              [57058012.82148681, 76831987.1785132]\n","35      Display > Organic Search > Referral  ...  [7391834.050253105, 2623250.57365777, 16384915...\n","38                       Display > Referral  ...             [33501108.56390939, 80708891.43609062]\n","39                           Organic Search  ...                                    [25418020000.0]\n","40             Organic Search > Paid Search  ...            [116152975.34855956, 165957024.6514404]\n","41  Organic Search > Paid Search > Referral  ...  [25725024.54588916, 43524361.64828748, 7424061...\n","44                Organic Search > Referral  ...              [674082087.187786, 2860957912.812214]\n","47                              Paid Search  ...                                     [2282920000.0]\n","48                   Paid Search > Referral  ...            [225600785.96289113, 463269214.0371089]\n","50                                 Referral  ...                                    [40676950000.0]\n","52                                   Social  ...                                       [61080000.0]\n","\n","[18 rows x 6 columns]"]},"metadata":{"tags":[]},"execution_count":12}]},{"cell_type":"markdown","metadata":{"id":"NvPNwnLruvUM","colab_type":"text"},"source":["- E os resultados agrupados por canais:"]},{"cell_type":"code","metadata":{"id":"si7ZVp1Suym8","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":220},"executionInfo":{"status":"ok","timestamp":1590964700709,"user_tz":180,"elapsed":1234,"user":{"displayName":"André Tocci Russi","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjUOmZJ5GKqIq-1s5NcPvDx4AUU0mOWUZw1W62r=s64","userId":"17444405581757145872"}},"outputId":"2e77230c-761d-4640-e583-f5f737ceb445"},"source":["shapley_results[1].reset_index()"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>channels</th>\n","      <th>value</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>Direct</td>\n","      <td>33042586811.941</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>Display</td>\n","      <td>549829572.671</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>Organic Search</td>\n","      <td>27222295557.146</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>Paid Search</td>\n","      <td>2718002172.263</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>Referral</td>\n","      <td>49310895885.980</td>\n","    </tr>\n","    <tr>\n","      <th>5</th>\n","      <td>Social</td>\n","      <td>61080000.000</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["         channels           value\n","0          Direct 33042586811.941\n","1         Display   549829572.671\n","2  Organic Search 27222295557.146\n","3     Paid Search  2718002172.263\n","4        Referral 49310895885.980\n","5          Social    61080000.000"]},"metadata":{"tags":[]},"execution_count":19}]},{"cell_type":"markdown","metadata":{"id":"PrQ1YpP52Qps","colab_type":"text"},"source":["Para rodar o modelo das **Cadeias de Markov**, basta aplicar o método .attribution_markov().\n","\n","No parâmetro de entrada transition_to_same_state, indica se irá ser retornado ou não a probabilidade de transição para o mesmo estado. Essa configuração **não afeta os resultados agregados** e que são atribuídos para cada canal, **mas sim os valores observados na matriz de transição**."]},{"cell_type":"code","metadata":{"id":"x08p7andppU9","colab_type":"code","colab":{}},"source":["markov_results = DP_tribution.attribution_markov(transition_to_same_state=False)"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"j8rJ5iGJ30vU","colab_type":"text"},"source":["O método retorna uma tupla contendo :\n","- Os resultados em nível de jornada, que também pode ser obeservado acessando o .DataFrame. "]},{"cell_type":"code","metadata":{"id":"7pw-1ZQP31MM","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":115},"executionInfo":{"status":"ok","timestamp":1590964557657,"user_tz":180,"elapsed":717,"user":{"displayName":"André Tocci Russi","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjUOmZJ5GKqIq-1s5NcPvDx4AUU0mOWUZw1W62r=s64","userId":"17444405581757145872"}},"outputId":"cc9cd82d-ab25-4c98-c5d3-59be866528e9"},"source":["markov_results[0].head()"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["0    [1.0]\n","1    [1.0]\n","2    [1.0]\n","3    [1.0]\n","4    [1.0]\n","Name: channelGrouping, dtype: object"]},"metadata":{"tags":[]},"execution_count":17}]},{"cell_type":"markdown","metadata":{"id":"VxPtJj2V4unf","colab_type":"text"},"source":["- E os resultados agrupados por canais, como retornado pelo modelo Shapley acima;"]},{"cell_type":"code","metadata":{"id":"U9r0nB1v4J-N","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":277},"executionInfo":{"status":"ok","timestamp":1590964678471,"user_tz":180,"elapsed":639,"user":{"displayName":"André Tocci Russi","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjUOmZJ5GKqIq-1s5NcPvDx4AUU0mOWUZw1W62r=s64","userId":"17444405581757145872"}},"outputId":"a4236df5-b056-4123-c396-2a5993e5d1c6"},"source":["markov_results[1]"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>channels</th>\n","      <th>attribution_markov_algorithmic</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>Display</td>\n","      <td>859295618.089</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>Affiliates</td>\n","      <td>0.000</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>(Other)</td>\n","      <td>0.000</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>Organic Search</td>\n","      <td>29002939438.611</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>Referral</td>\n","      <td>46028648315.510</td>\n","    </tr>\n","    <tr>\n","      <th>5</th>\n","      <td>Social</td>\n","      <td>52278245.740</td>\n","    </tr>\n","    <tr>\n","      <th>6</th>\n","      <td>Direct</td>\n","      <td>33804951171.975</td>\n","    </tr>\n","    <tr>\n","      <th>7</th>\n","      <td>Paid Search</td>\n","      <td>3156577210.076</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["         channels  attribution_markov_algorithmic\n","0         Display                   859295618.089\n","1      Affiliates                           0.000\n","2         (Other)                           0.000\n","3  Organic Search                 29002939438.611\n","4        Referral                 46028648315.510\n","5          Social                    52278245.740\n","6          Direct                 33804951171.975\n","7     Paid Search                  3156577210.076"]},"metadata":{"tags":[]},"execution_count":18}]},{"cell_type":"markdown","metadata":{"id":"Dy66fGh_2ZUt","colab_type":"text"},"source":["- Matrix de transição:"]},{"cell_type":"code","metadata":{"id":"mNrja1mO2Zc8","colab_type":"code","colab":{}},"source":["ax, fig = plt.subplots(figsize=(15,10))\n","sns.heatmap(markov_results[2].round(3), cmap=\"YlGnBu\", annot=True, linewidths=.5)"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"b0LlWef949Rm","colab_type":"text"},"source":["- **Removal Effect**, quarto output dos resultados attribution_markov, que como detalhado anteriormente, nos auxilia a entender a impertância de cada canal analisado:"]},{"cell_type":"code","metadata":{"id":"1dPyfO8K49YP","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":332},"executionInfo":{"status":"ok","timestamp":1590964911054,"user_tz":180,"elapsed":849,"user":{"displayName":"André Tocci Russi","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjUOmZJ5GKqIq-1s5NcPvDx4AUU0mOWUZw1W62r=s64","userId":"17444405581757145872"}},"outputId":"74562fda-6cf6-4dfa-e449-d6af7a600ea1"},"source":["ax, fig = plt.subplots(figsize=(2,5))\n","sns.heatmap(markov_results[3].round(3), cmap=\"YlGnBu\", annot=True, linewidths=.5)"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.axes._subplots.AxesSubplot at 0x7f2218cdc6d8>"]},"metadata":{"tags":[]},"execution_count":22},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 144x360 with 2 Axes>"]},"metadata":{"tags":[]}}]},{"cell_type":"markdown","metadata":{"id":"0pvB6Nr3s7QY","colab_type":"text"},"source":["#### Comparando os Resultados\n","\n","Agora que rodamos ambos os modelos, podemos comparar seus resultados acessando o atributo group_by_channels_models, que retorna todos os diferentes modelos aplicados nos passos anteriores."]},{"cell_type":"code","metadata":{"id":"8UFrFSzus-dU","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":277},"executionInfo":{"status":"ok","timestamp":1593454835988,"user_tz":180,"elapsed":543,"user":{"displayName":"André Tocci Russi","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjUOmZJ5GKqIq-1s5NcPvDx4AUU0mOWUZw1W62r=s64","userId":"17444405581757145872"}},"outputId":"05e071e2-cb13-46c3-b2da-5f24c49afe92"},"source":["DP_tribution.group_by_channels_models"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/html":["<div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>channels</th>\n","      <th>attribution_shapley_size4_conv_rate_algorithmic</th>\n","      <th>attribution_first_click_heuristic</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>Direct</td>\n","      <td>33042586811.941</td>\n","      <td>37807280000.000</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>Display</td>\n","      <td>549829572.671</td>\n","      <td>391590000.000</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>Organic Search</td>\n","      <td>27222295557.146</td>\n","      <td>27977540000.000</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>Paid Search</td>\n","      <td>2718002172.263</td>\n","      <td>2488230000.000</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>Referral</td>\n","      <td>49310895885.980</td>\n","      <td>44178970000.000</td>\n","    </tr>\n","    <tr>\n","      <th>5</th>\n","      <td>Social</td>\n","      <td>61080000.000</td>\n","      <td>61080000.000</td>\n","    </tr>\n","    <tr>\n","      <th>6</th>\n","      <td>(Other)</td>\n","      <td>0.000</td>\n","      <td>0.000</td>\n","    </tr>\n","    <tr>\n","      <th>7</th>\n","      <td>Affiliates</td>\n","      <td>0.000</td>\n","      <td>0.000</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>"],"text/plain":["         channels  ...  attribution_first_click_heuristic\n","0          Direct  ...                    37807280000.000\n","1         Display  ...                      391590000.000\n","2  Organic Search  ...                    27977540000.000\n","3     Paid Search  ...                     2488230000.000\n","4        Referral  ...                    44178970000.000\n","5          Social  ...                       61080000.000\n","6         (Other)  ...                              0.000\n","7      Affiliates  ...                              0.000\n","\n","[8 rows x 3 columns]"]},"metadata":{"tags":[]},"execution_count":13}]},{"cell_type":"code","metadata":{"id":"Zq09OuvVqSvO","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":358},"executionInfo":{"status":"ok","timestamp":1593454914175,"user_tz":180,"elapsed":1019,"user":{"displayName":"André Tocci Russi","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjUOmZJ5GKqIq-1s5NcPvDx4AUU0mOWUZw1W62r=s64","userId":"17444405581757145872"}},"outputId":"cedb47c0-f62e-4ffb-f64d-752336111ee6"},"source":["DP_tribution.group_by_channels_models.sum()\n","\n","DP_tribution.attribution_last_click()"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(0             [0.0]\n"," 1             [0.0]\n"," 2             [0.0]\n"," 3             [0.0]\n"," 4             [0.0]\n","             ...    \n"," 57701         [0.0]\n"," 57702         [0.0]\n"," 57703         [0.0]\n"," 57704         [0.0]\n"," 57705    [0.0, 0.0]\n"," Length: 57706, dtype: object,\n","          channels  attribution_last_click_heuristic\n"," 0         (Other)                             0.000\n"," 1      Affiliates                             0.000\n"," 2          Direct                   30260240000.000\n"," 3         Display                     817090000.000\n"," 4  Organic Search                   29853940000.000\n"," 5     Paid Search                    2809400000.000\n"," 6        Referral                   49102940000.000\n"," 7          Social                      61080000.000)"]},"metadata":{"tags":[]},"execution_count":16}]},{"cell_type":"code","metadata":{"id":"CJHui3UDptrc","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":483},"executionInfo":{"status":"ok","timestamp":1593454918626,"user_tz":180,"elapsed":815,"user":{"displayName":"André Tocci Russi","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GjUOmZJ5GKqIq-1s5NcPvDx4AUU0mOWUZw1W62r=s64","userId":"17444405581757145872"}},"outputId":"85adc4ef-1c17-4a4f-c670-f7b2b5e6676d"},"source":[" DP_tribution.plot(model_type='heuristic')"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.axes._subplots.AxesSubplot at 0x7ff1f3bf11d0>"]},"metadata":{"tags":[]},"execution_count":17},{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAABIEAAAHCCAYAAAB4wtmkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzde7xVdZ0//teBA8pNAwUUbwjq0tIoMc2mC9Y3bUyaJhEN0+5kpd8k81f5VTerDEcdmy5mRpdJnCxvqaM5lY1pNTqWTqmZLk3xksRFlIsgcjnn98c+NISAaGwOsJ7Px4MH5+y19/q8P+ess/bar/X5rNXW2dkZAAAAALZsPbq7AAAAAABaTwgEAAAAUANCIAAAAIAaEAIBAAAA1IAQCAAAAKAGhEAAAAAANdDe3QUURbFvkmuT/EtVVRes43kDk3w/yTNVVY3reqxXku8m2S3JiiTvr6rq4ZYXDQAAALCZ6daRQEVR9Evy1ST/uR5PvyjJr1Z7bEKSeVVVvT7JF5KcvWErBAAAANgydPdIoOeSHJ7k0ysfKIri5UkuSNKZZGGS91VVNS/Jh5KMTvKqVV7/liTTur7+WZLvbISaAQAAADY73ToSqKqq5VVVPbvaw19N8pGqqt6S5KdJPt713IVrWMUOSeZ0Le9I0lkURe8WlgwAAACwWerukUBrcmCSbxZFkSRbJfnNi3htW0sqAgAAANjMbYoh0OIkh1RV1bkez52R5migu7ouEt1WVdXSllYHAAAAsBnaFG8Rf1eStyVJURTHFEXxlnU896dJjur6emySn7e4NgAAAIDNUltn5/oMuGmNoihGJzk/yfAky5I8keT/JfmnJB1Jnk3zDmDz07yD2MuS7JTk3iSfS3JLkm8l2TPNi0y/r6qqxzdqJwAAAAA2A90aAgEAAACwcWyK08EAAAAA2MCEQAAAAAA1IAQCAAAAqAEhEAAAAEANCIEAAAAAakAIBAAAAFAD7d1dwKruvPPOIe3t7d9Ksm8EVGx6OpL8fvny5R8aPXr07O4uBgAAAF6MTSoEam9v/9YOO+ywz+DBg5/u0aNHZ3fXA6vq6OhomzNnzstnzpz5rSTv6O56AAAA4MXY1Ebb7Dt48OAFAiA2RT169OgcPHjw/DRHqgEAAMBmZVMLgXoIgNiUdW2fm9rfDQAAALyglk0HK4piTJIrktzb9dA9VVWd1Kr2AAAAAFi7Vl8T6Jaqqsa91BePPnXa6A1ZzJ3nHX/nhlwfAAAAwObCtJb1cMstt/SfNWtWe5Jcd9112z733HNtqz9n4sSJIxYvXtx20kknDb/++uu3XZ/1Pvroo71vv/32vkly2mmn7fLwww/33hD1/uAHPxj05je/ed9f/OIX/d/73veOXJ/XzJ8/v8dPf/rTbV5sWwcccMCoZN31n3322cMuuuiiweuzvnHjxhX33HPP1i+2jrVZ+XtZ07JW/fwBAABgU9TqkUAvL4ri35MMSlJWVXXjygVlWU5OkkajMbksyweSHPGOd7yj9+zZs/cZMmTIffPmzdt5QxezaNGiQUuWLNl2u+22m/7kk0/u0bdv3zlbb731wlmzZo3acccdf/vMM89sv2zZsn4DBw58dM6cOcWAAQNmtLe3L7nqqqv2/PCHP3xfnz59tps2bdrQMWPG/H7+/Pl7b7vtto/26NFjxVNPPbXn1KlTfz9//vxhSfomeXrWrFn7Dho06MGOjo6e8+fP321ln3r06LFsm222mTVz5sxX3nzzzTMXL1681YgRI3aZMmVK9fTTT+/2zDPPLOrfv/+Tf/7zn189dOjQu5YsWTJg8eLFg7fffvs/zp07d/ett956fr9+/Z6aMWPG6GHDht25pj7deuut25xwwgm93vjGNz6z//77b/3000/vtnqfnnzyyX122GGHuxcsWDC0o6Oj1+9+97v5v/zlL3d/05ve9ODKPg0dOnRln7LtttvOWFOf2traei5YsGDolClTHp85c+Yrly9fft/y5cu3Xrhw4bDBgwdXTz/99G6dnZ29ksxfnz4leWbp0qXbzp07d8cX+3tavU8ve9nL/nTWWWdt1bNnzz7Lli17Xp9uueWW5xYuXLjz/vvvf3+j0Zg7f/78kUme93vafvvt/6pPS5cu3a4sy4mNRmNqWZYLkwxLMibJxEajMbYsy0uTXN9oNC4ty7Kz0Wi0lWU5IckRjUZjQlmW1yWZmuTmJDMajcaAsiwnJjmg0WhMLMvy5iSTkzyQ5I5GozGsLMtTkgxrNBqnlGV5Z5KJSRZ2tbPXmv6ekgxIMrXRaIwuy/L8rrbOL8tyRpIDkuyVZHKj0RhTluXUrrb0SZ/0SZ/0SZ/0SZ/0SZ/0SZ/0aTPuU9ahrbOzNddhLopipySvT3J5khFJfp5kj6qqlq7tNXfdddcjo0aNenLl990xHWz+/Pk9TjzxxBFLlizpsWTJkh6HHHLIvEsuuWSHnXba6bmjjz569jnnnLPbXnvttejss89+5DOf+czwPn36dEyYMGH2Oeecs+sNN9xw72mnnbZrr169Op944omt5s2b1z5lypRHBg0atPzEE08c+aMf/ei+JHn729++T6PRePSUU04Z2bNnz85TTjnl8YsvvniHRqPx6C677LJs0qRJw5955pmey5cvbzv99NMfHz169OIxY8bs+4Y3vGHe3Xff3b9///4rpk2b9mDPnj2fV/+NN964zemnn777tttuu+y000577FOf+tTIO+64465x48YVI0aMeDZJjjnmmDllWe7Wq1evzl69enVceOGFD48fP37vxYsX9/zABz4w44Mf/OCTz1txku9973uDvv/97w9ta2vrPO6442aNHz/+6QMOOGDUyvU3Go1Hd95552UnnXTS7osXL+7Zr1+/FRdeeOHDF1xwwQ4DBw5cdsIJJ8z56Ec/uvvrX//6+ccee+xTa2pj3Lhxxatf/eqFd911V/8FCxa0f+Mb3/jjbrvttvTzn//8Tr/73e/6r1ixou3d73737KOPPvqpk046afhhhx329BFHHDH/+uuv3/YnP/nJwE9+8pMzPvnJT+6++u/ltttu6//Vr351p969e3cMGjRo2Re+8IXHjjzyyH1W//mvqf4BAwZ0rFrjXXfdtf2oUaOGv/AWBwAAAJuOlk0Hq6rqiaqqLquqqrOqqoeSzEyyU6va21BmzpzZa9y4cU9eccUVD0yaNOmJe++9t9/IkSOfPfvss6cfd9xxcwcOHLjs4osvfrB3796dDz30UN+vfOUrDx9xxBHzV11HW1tb52WXXfbASSed9MSFF16445raGTx48PLDDz987jHHHDNr7Nixf3n9RRddNGS//fZbdMUVVzxw2mmnPT5lypRduuraaty4cXOvvfba+xcuXNjz7rvv7rOm9b71rW9dcOCBB87/v//3/z4xZsyYZ1Zdtueeez577rnnPnb55ZdvP378+NlXXnll9eEPf3jmzJkzex133HEz3/zmNz+1tgBowYIFPb75zW8Ou/zyy++fNm3agz/60Y8Grel5F1xwwdDXve51C374wx9WBx544IKbbrrpL1PMvvzlLw/dcccdl64tAFppu+22W3b55Zc/cPDBB8+/7rrrXvaLX/yi/5///OfeV111VfWDH/yg+ta3vrXj2qZ4Jcmafi/f+973hpx66qmPX3XVVdXhhx/+1IoVK7Kmn/+66gcAAIDNWSvvDnZskh2rqvrnoih2SDI0yROtam9DGTp06PIvf/nLA7/73e8OXbZsWY+tt966Y23P3XHHHZ/bfvvtV6z++EEHHbQwSV7zmtcs+vKXv/yiprXdd999/T760Y/+uev1i2fMmLFVkvTp02fFqFGjnk2SwYMHL50/f/7zhwG9gP33339Rkrz1rW+dd9ZZZ+02ffr0rf/hH/7hqX322WfJb37zm34vUNfWu+6665K+fft29u3bd8XFF1/80JqeV1VV35NPPnlGkpx00kmzk+QPf/hD39tvv32bWbNm9b7++uvve6E6DzrooGeSZOjQoUvnzZvXfscdd/T/wx/+0G/cuHFFknR0dOTPf/5zr7W9fk2/l0MPPfSpz3/+87sddthhc9/1rnc9teOOOy5f3/oBAABgS9DKC0P/e5I3FUXxyyTXJvnouqaCbSouuuiiIUOGDFl69dVXV41G49F1Pbe9vX2Nc+na2tpW/bpz1e+TZPny5WsdxZKkc9Upeh0dHW1J0rNnz9XbWtc61qh3794dSfKWt7xl4dVXX33fyJEjl3z605/e/ec///mAF3ptz54909Gx1jzsL3r06LHG582fP7+9d+/enb/61a/6v9A6Vv25dnZ2tvXq1atj7NixT1555ZXVlVdeWf3nf/7nvSNHjvyrbWnVn+mafi/HHnvsU//2b/9WDRw4cPlHPvKRPe+77741Xnx6bfUDAADA5q5lI4GqqlqYZOzfso7uuKX7008/3V4UxbNJ8uMf/3jg8uXL23r37t2xMmRoa2t7oRAnd9xxR/+jjjrq6V//+tf9dttttyXbbLPNinnz5vXq6OjIrFmz2mfOnLlV17o6V1/XK17xisW33nrrgIMPPnjRbbfd1m/48OHPbug+XnTRRYMPPfTQ+e9+97uf6uzszO9///u+AwcOXL5ixYq19mvvvfde8qc//WnrhQsX9mhvb+98//vfv8ell1764OrPe/nLX77oV7/61TYHHnjg4m9961vbb7XVVp1JcyTOm970poWf+MQnRlxzzTX39e3bd70vRrX//vsvOu+883Y5+eSTZy5durRt8uTJO59zzjmP9+/ff8Xs2bN7Jc2f+brWcc455+z44Q9/ePYHP/jBJ+fOndvr/vvv33pNP/811X/cccfNXd9aAQAAYFPV6ruDbXaOPPLIuZ/97Gd3v/HGGwdOmDBh9o033jjo8MMPf3LSpEkjL7jggj++6lWvWjh+/Pi9p0yZMn1d6zn++OP3mD17du9zzz13+qBBg1bsv//+C8aOHbvPHnvssXjkyJGLk2T06NGLzjjjjOHbbbfdX6YmnXDCCbMmTZo0fPz48Xt1dHS0TZ48eZ2jkV6K4cOHP3fSSSeN7Nev34revXt3/PM///Mjc+bMab/gggt2+vKXv7z0E5/4xKzVX9O/f/+OE044Ycaxxx67V5K85z3vmdWjx/MHkp1wwgmzP/GJTwwfN25c0bdv3xUXXHDBw1/72td2SJJ99tlnyWGHHfbUWWedtfOUKVMeX996/+7v/m7RTTfdtPCd73zn3knajjrqqNlJ8q53vWvupz/96RE/+9nPBu61116L17WOYcOGLX3Pe96z14ABA1YMGDBg+cc//vFZ/fv371jDz/959a9vnQBQN499br/W3GFkE7Hrmfe86JHXALApa9ndwV6K1e8OBpsidwcDgCYhEABsXowE2kzdfvvtfc8777xdVn/80EMPfWrixIlzXup6r7vuum0vvvjiHVZ//D3vec+sd77znfNe6npX1araAQAAgLUzEgheJCOBAKDJSCAA2Ly08u5gAAAAAGwihEAAAAAANSAEAgAAAKiBTfrC0I99br/RG3J9u555z50bcn0AAAAAmwsjgdbDLbfc0n/WrFntSfPuWc8999zzLhI4ceLEEYsXL2476aSThl9//fXbrs96H3300d6333573yQ57bTTdnn44Yd7b4h63/CGN+y3cOHCF/W7veqqqwa+2HZW9vXHP/7xNlOnTh28pudMnz6999vf/vZ91md906ZN2+6MM87Y+cXWsa71XXPNNS9b2/KVfV5X/QAAALClEAKthyuvvHL72bNntyfJxRdfvMPSpUufFwJNnTr14b59+76oO2TccsstA37729/2S5IpU6Y8PmLEiKUbpuIX7zvf+c7zbgu/vt72trct2BRv7X788cfPXdtt7Z977rm2Sy65ZGiy6dYPAAAAG9ImPR2sO8yfP7/HiSeeOGLJkiU9lixZ0uOQQw6Z96tf/epl06dP73P00UfPvv/++/u9973v3fPss89+5DOf+czwPn36dEyYMGH2Oeecs+sNN9xwb5LcdNNNL7vkkkuGzps3r33KlCmPDBo0aPmJJ5448kc/+tF9SfL2t799n0aj8eg3v/nNYT179uzcaaedll588cU7NBqNR3fZZZdlkyZNGv7MM8/0XL58edvpp5/++OjRoxePGTNm3ze84Q3z7r777v79+/dfMW3atAd79uy5zr787ne/61OW5a7t7e2dbW1t+frXv/7QgAEDOk488cTd586d22vp0qVtH//4x2fcf//9faZPn97nAx/4wMjvfOc7D61pXY888kjvT33qU7t3dHRkhx12WPqVr3xl+spl06ZN2+7BBx/s8/nPf/5P//Iv/zL0pptuGtTW1tY5adKkJ4YPH/7cyuf9x3/8xzbf+973hnz3u9/9Y3v7mje9OXPm9PrgBz848pFHHtn6+OOPn/ne97537i9+8Yv+X/rSl3Zqb2/vHDJkyNLzzz//0dtuu63/JZdcMvjb3/72w0lywAEHjLrjjjvuGjduXDFixIhnk2TgwIHLBw4cuOyYY4556qMf/eiIZcuW9Vi2bFnbmWee+dgPfvCD7adPn97n1FNP3XW//fZbtLb6DznkkIUvbgsCAACATZORQKuZOXNmr3Hjxj15xRVXPDBp0qQn7r333n4jR4589uyzz55+3HHHzR04cOCyiy+++MHevXt3PvTQQ32/8pWvPHzEEUfMX3UdbW1tnZdddtkDJ5100hMXXnjhjmtqZ/DgwcsPP/zwucccc8yssWPH/uX1F1100ZD99ttv0RVXXPHAaaed9viUKVN26aprq3Hjxs299tpr71+4cGHPu+++u88L9WXOnDm9zjjjjMeuuOKKB0aNGvXM5Zdfvt0999zTZ968ee1XXXVVNW3atAfnzZvXfvLJJ8/q27fvirUFQEly7rnn7vTe97535g9/+MNq++23X3bHHXf0W/05DzzwwFY33XTToGuuuea+888/f/q11147aOWyBx98cKtvfOMbwy644ILpawuAkmTGjBlbfeMb33joa1/72h+///3vD02Ss88+e9dvfvObf7z88ssfGDRo0PIf/vCH65y6tueeez577rnnPrby+5///OcDhgwZsvTKK6+svvjFLz48Z86c9o997GMzd9555yXnnXfeX563rvoBAABgcycEWs3QoUOX/+QnPxn4j//4j8UXv/jFnefPn7/WxGLHHXd8bvvtt1+x+uMHHXTQwiR5zWtes+ixxx7b+sW0f9999/V73etet/L1i2fMmLFVkvTp02fFqFGjnk2SwYMHL50/f/66hwElGTJkyLLzzz9/5yOPPLK48cYbB82bN6997733XvLss8/2/NjHPrb7zTffPOCoo456an3qeuCBB/oefPDBzyTJ5z73uT+99rWvXbT6c+66666+L3/5y5/p2bNn9txzz+e+9KUvPZokS5Ys6fGxj31sjzPOOOOxl73sZc/7ea1q3333XdTe3p6ddtpp2aJFi3rOnDmzfcaMGVt95CMf2WPcuHHF//zP/wyYOXNmr3WtY//99/+r2g466KBF9957b/9TTjll14ceemjrt73tbQvW9Lq11Q8AAABbAiHQai666KIhQ4YMWXr11VdXjUZjnSFAe3v7Gq8B1NbWturXnat+nyTLly9/3jWFVtHZ2fm/q+3o6GhLkp49e67e1rrWkSQ566yzdjnuuONmXXXVVdU//uM/zkmSfv36dVxzzTX3vfvd755zyy23bDtp0qThL7SeJOnRo0fnylrWpmfPnmt8zpNPPtl71KhRCy+55JIXvPjy6v3s3bt353bbbbfsyiuvrK688srq+uuvv2/SpEmzkvzV81asWNG2yms6Vl02bNiwZTfccMMfDjvssHnf//73B//TP/3TGkdnra1+AAAA2BJs0tcE6o5buj/99NPtRVE8myQ//vGPBy5fvrytd+/eHSuDm7a2thcKcXLHHXf0P+qoo57+9a9/3W+33XZbss0226yYN29er46OjsyaNat95syZW3Wtq3P1db3iFa9YfOuttw44+OCDF9122239hg8f/uxL7cv8+fPbR4wY8dySJUva/uu//mvbfffdd9Gdd97Zt6qqrSdMmPDUQQcdtGjcuHFFknR2dq6zT3vvvffim2++ecD48eOf/sIXvjBs5WinVb361a9ePHXq1GHLli3L3Llz2z/72c/uduaZZz6+8847Lzn33HMfO/roo/f66U9/us2hhx66xpE4azJo0KAVSXLvvfdu/YpXvGLJ17/+9SGve93rFm6zzTYr5s6d2ztJ7rrrrj7PPvvsWkdG/exnPxuwbNmytr//+79fsM8++zx7+umn79ajR4+5qwZHa6v/X//1X9c6RQ4AAAA2J5t0CNQdjjzyyLmf/exnd7/xxhsHTpgwYfaNN9446PDDD39y0qRJIy+44II/vupVr1o4fvz4vadMmTJ9Xes5/vjj95g9e3bvc889d/qgQYNW7L///gvGjh27zx577LF45MiRi5Nk9OjRi84444zh22233fKVrzvhhBNmTZo0afj48eP36ujoaJs8efJLnpJ09NFHz/74xz++x7Bhw55befHqN77xjfOvv/767a688srBPXr0yPve976ZSTJy5MjFY8eO3ee66667b03r+tSnPvXEqaeeuvtll102ZOjQoUvHjBkz49prr91u1efsvvvuSw8//PC5Rx111N6dnZ05+eSTn1i5rEePHjn77LMfPeGEE/Z47Wtfe98222zT8fxW1qwsy0c+85nPDO/Vq1fn9ttvv+z973//nF69enX26dNnxTvf+c69X/nKVz4zZMiQtd5ZbeTIkc998pOf3P3b3/72jj169Og88cQTZ+y4447Lli9f3vahD31oxBvf+Mb5L1Q/AAAAbO7aVp161N3uuuuuR0aNGvVkd9cB63LXXXdtP2rUqOHdXQcAdLfHPrffpnMg2QK7nnmPaeIAbFGMBNpM3X777X3PO++8XVZ//NBDD31q4sSJc17KOp977rm2Y489dq/VH991112XfPGLX9xgF0n+wAc+MHLBggV/te3169dv+cUXX2zqFQAAALTIpjYS6OH99tvv6R49emw6RcEqOjo62u65556Bo0aNGtHdtQBAdzMSCAA2L5va3cF+P2fOnG3doYlNUUdHR9ucOXO2TfL77q4FAAAAXqxNajrY8uXLPzRz5sxvzZw5c99segEVdCT5/fLlyz/U3YUAAADAi7VJhUCjR4+eneQd3V0HAAAAwJbGaBsAAACAGhACAQAAANSAEAgAAACgBoRAAAAAADUgBAIAAACoASEQAAAAQA0IgQAAAABqQAgEAAAAUANCIAAAAIAaEAIBAAAA1IAQCAAAAKAGhEAAAAAANSAEAgAAAKgBIRAAAABADQiBAAAAAGqgvbsLAADYUo0+dVpnd9fQSlcP6O4KAIAXw0ggAAAAgBoQAgEAAADUgBAIAAAAoAaEQAAAAAA1IAQCAAAAqAEhEAAAAEANCIEAAAAAakAIBAAAAFADQiAAAACAGhACAQAAANSAEAgAAACgBoRAAAAAADUgBAIAAACoASEQAAAAQA0IgQAAAABqQAgEAAAAUANCIAAAAIAaEAIBAAAA1IAQCAAAAKAGhEAAAAAANSAEAgAAAKgBIRAAAABADQiBAAAAAGpACAQAAABQA0IgAAAAgBoQAgEAAADUgBAIAAAAoAZaGgIVRdGnKIqHiqJ4XyvbAQAAAGDdWj0S6PQkT7W4DQAAAABeQMtCoKIo9k7y8iQ/alUbAAAAAKyf9hau+/wkJyZ575oWlmU5OUkajcbksiwfSHJEkgFJpjYajdFlWZ6fZEaj0Ti/LMsZSQ5IsleSyY1GY0xZllOT3NFoNKaWZbkwybAkY5JMbDQaY8uyvDTJ9Y1G49KyLDsbjUZbWZYTkhzRaDQmlGV5XZKpSW7uamdAWZYTkxzQaDQmlmV5c5LJSR7oamdYWZanJBnWaDROKcvyziQTkyzsamcvfdInfdInfdInfdKnv+7T7mHztXlve1vi35M+6ZM+6ZM+rU+fsg5tnZ2d61r+khRFcXySXauqOqsoislJHqmq6rsbvCEAgE3Y6FOnbfgDrU3I1QPO6+4SWmrXM+9p6+4aAGBDatVIoLcnGVEUxRFJdk7yXFEUf6qq6mctag8AAACAdWhJCFRV1dErv15lJJAACAAAAKCbtPruYAAAAABsAlp5YegkSVVVk1vdBgAAAADrZiQQAAAAQA0IgQAAAABqQAgEAAAAUANCIAAAAIAaEAIBAAAA1IAQCAAAAKAGhEAAAAAANSAEAgAAAKgBIRAAAABADQiBAAAAAGqgvbsLeLFGnzqts7traKU7zzu+rbtrAAAAALY8RgIBAAAA1IAQCAAAAKAGhEAAAAAANSAEAgAAAKgBIRAAAABADQiBAAAAAGpACAQAAABQA0IgAAAAgBoQAgEAAADUgBAIAAAAoAaEQAAAAAA1IAQCAAAAqAEhEAAAAEANCIEAAAAAakAIBAAAAFADQiAAAACAGhACAQAAANSAEAgAAACgBoRAAAAAADUgBAIAAACoASEQAAAAQA0IgQAAAABqQAgEAAAAUANCIAAAAIAaEAIBAAAA1IAQCAAAAKAGhEAAAAAANSAEAgAAAKgBIRAAAABADQiBAAAAAGpACAQAAABQA0IgAAAAgBoQAgEAAADUgBAIAAAAoAaEQAAAAAA1IAQCAAAAqAEhEAAAAEANCIEAAAAAakAIBAAAAFADQiAAAACAGhACAQAAANSAEAgAAACgBoRAAAAAADUgBAIAAACoASEQAAAAQA0IgQAAAABqQAgEAAAAUANCIAAAAIAaEAIBAAAA1IAQCAAAAKAGhEAAAAAANSAEAgAAAKgBIRAAAABADbS3asVFUfRN8t0kQ5NsneTzVVVd36r2AAAAAFi7loVAScYmuaOqqnOLotgtyY1JhEAv4LHP7dfZ3TW00q5n3tPW3TUAAABAHbUsBKqq6rJVvt0lyZ9a1RYAAAAA69bKkUBJkqIobk2yc5IjVn28LMvJSdJoNCaXZflA1/IBSaY2Go3RZVmen2RGo9E4vyzLGUkOSLJXsnurS6aFyrIckGRMkomNRmNsWZaXJoaemF0AACAASURBVLm+0WhcWpZlZ6PRaCvLckKSIxqNxoSyLK9LMjXJzWluDwPKspyY5IBGozGxLMubk0xO8kCSOxqNxrCyLE9JMqzRaJxSluWdSSYmWdjVzl4vfdvL5EajMaYsy6ldbU0ty3JhkmH6pE/6pE/6pE9r7pPjls3Z5r3tbYl/T/qkT/qkT/q0Pn3KOrR1drZ+9lFRFK9KMi3JqKqq/qYGR586bYueLnX1gPO6u4SWMh0MuoepptA9HLds3uxbANjStOzuYEVRjC6KYpckqarqd2mOOhrcqvYAAAAAWLtW3iL+jUlOSZKiKIYm6Z/kyRa2BwAAAMBatDIEuijJkKIofpnkR0k+XlVVRwvbAwAAAGAtWnl3sGeTTGjV+gEAAABYf60cCQQAAADAJkIIBAAAAFADQiAAAACAGhACAQAAANSAEAgAAACgBoRAAAAAADUgBAIAAACoASEQAAAAQA0IgQAAAABqoL27CwDYFIw+dVpnd9fQSlcP6O4KAACA7mYkEAAAAEANCIEAAAAAakAIBAAAAFADQiAAAACAGhACAQAAANSAEAgAAACgBtYrBCqKYueiKF7f9fVWrS0JAAAAgA3tBUOgoigmJbksyde6HjqnKIpPt7QqAAAAADao9RkJ9M6qqv4uyVNd309K8s7WlQQAAADAhrY+IVDPrv87u/7fOkl7a8oBAAAAoBXWJwS6tCiKm5LsWRTF15P8Nsl3WlsWAAAAABvSC47oqarqwqIobkhyYJLnkkypqurxllcGAAAAwAbzgiFQURT/mv+dCpYk/1AURaqq+kDrygIAAABgQ1qfa/tcucrXvZK8PsnS1pQDAAAAQCusz3SwH6320DVd08MAAAAA2Eysz3Sww1d7aMckI1pTDgAAAACtsD7TwY5a5evOJAuSHNuacgAAAABohbWGQEVR9O368uMbqRYAAAAAWmRdI4HuzV/fFWyltq7HTQkDAAAA2EysNQSqqmr3tS0riuJ9LakGAAAAgJZYnwtDH5Dk00m263qod5Idkny3dWUBAAAAsCH1WI/nfDXJhUn6Jzk1yc1JTm5hTQAAAABsYOsTAi2uqurnSZ6rqurOqqpOT3Jii+sCAAAAYANan1vELyqK4h1JphdFMSXJQ0l2bW1ZAAAAAGxI6zMSaOckr0xyXpIlXV8f38qiAAAAANiw1icEemeS+Un+Jck7ksxJ8kwriwIAAABgw3rB6WBVVT2W5sWhv1oUxc5JvpDk7jTvEgYAAADAZmB9bhG/c5KxXf+GJbkhyetaXBcAAAAAG9D6XBj62iQ/TPKpqqr+0OJ6AAAAAGiB9ZkONnpjFAIAAABA66zPhaEBAAAA2MwJgQAAAABqQAgEAAAAUANCIAAAAIAaEAIBAAAA1IAQCAAAAKAGhEAAAAAANSAEAgAAAKgBIRAAAABADQiBAAAAAGpACAQAAABQA0IgAAAAgBoQAgEAAADUgBAIAAAAoAaEQAAAAAA1IAQCAAAAqAEhEAAAAEANCIEAAAAAakAIBAAAAFADQiAAAACAGhACAQAAANSAEAgAAACgBoRAAAAAADUgBAIAAACoASEQAAAAQA20t3LlRVGcm+QNXe2cXVXVD1vZHgAAAABr1rKRQEVRHJJk36qqDk7ytiRfalVbAAAAAKxbK6eD/SLJUV1fz0vSryiKni1sDwAAAIC1aNl0sKqqViRZ1PXtB5Pc0PVYkqQsy8lJ0mg0Jpdl+UCSI5IMSDK10WiMLsvy/CQzGo3G+WVZzkhyQJK9kt1bVTIbQVmWA5KMSTKx0WiMLcvy0iTXNxqNS8uy7Gw0Gm1lWU5IckSj0ZhQluV1SaYmuTnN7WFAWZYTkxzQaDQmlmV5c5LJSR5Ickej0RhWluUpSYY1Go1TyrK8M8nEJAu72tnrpW97mdxoNMaUZTm1q62pZVkuTDJMnzb/Ptm3bN669i2b5ba3Jf496dOqfbJv2Zxt3tvelvj3pE/6pE/6pE/r06esQ1tnZ+e6lv/NiqL4hySnJTm0qqr5f+v6Rp86rbUFd7OrB5zX3SW01K5n3tPW3TXAmti3bN7sW9hU2bds3uxbANjStPrC0Icl+X9J3rYhAiAAAAAAXpqWhUBFUWyb5Lwk/6eqqqda1Q4AAAAAL6yVI4GOTrJ9ksuLolj52PFVVT3WwjYBAAAAWINWXhh6apoXOgIAAACgm7XyFvEAAAAAbCKEQAAAAAA1IAQCAAAAqAEhEAAAAEANCIEAAAAAakAIBAAAAFADQiAAAACAGhACAQAAANSAEAgAAACgBoRAAAAAADUgBAIAAACoASEQAAAAQA0IgQAAAABqQAgEAAAAUANCIAAAAIAaEAIBAAAA1IAQCAAAAKAGhEAAAAAANSAEAgAAAKgBIRAAAABADQiBAAAAAGpACAQAAABQA0IgAAAAgBoQAgEAAADUgBAIAAAAoAaEQAAAAAA1IAQCAAAAqAEhEAAAAEANCIEAAAAAakAIBAAAAFADQiAAAACAGhACAQAAANSAEAgAAACgBoRAAAAAADUgBAIAAACoASEQAAAAQA0IgQAAAABqQAgEAAAAUANCIAAAAIAaEAIBAAAA1IAQCAAAAKAGhEAAAAAANSAEAgAAAKgBIRAAAABADQiBAAAAAGpACAQAAABQA0IgAAAAgBoQAgEAAADUgBAIAAAAoAaEQAAAAAA1IAQCAAAAqAEhEAAAAEANCIEAAAAAakAIBAAAAFADQiAAAACAGhACAQAAANSAEAgAAACgBoRAAAAAADUgBAIAAACoASEQAAAAQA0IgQAAAABqQAgEAAAAUANCIAAAAIAaEAIBAAAA1IAQCAAAAKAGWhoCFUWxb1EUDxVFcWIr2wEAAABg3VoWAhVF0S/JV5P8Z6vaAAAAAGD9tHIk0HNJDk8yo4VtAAAAALAe2lu14qqqlidZXhTFGpeXZTk5SRqNxuSyLB9IckSSAUmmNhqN0WVZnp9kRqPROL8syxlJDkiyV7J7q0pmIyjLckCSMUkmNhqNsWVZXprk+kajcWlZlp2NRqOtLMsJSY5oNBoTyrK8LsnUJDenuT0MKMtyYpIDGo3GxLIsb04yOckDSe5oNBrDyrI8JcmwRqNxSlmWdyaZmGRhVzt7vfRtL5MbjcaYsiyndrU1tSzLhUmG6dPm3yf7ls1b175ls9z2tsS/J31atU/2LZuzzXvb2xL/nvRJn/RJn/RpffqUdWjr7Oxc1/K/WVEUk5M8WVXVBRtifaNPndbagrvZ1QPO6+4SWmrXM+9p6+4aYE3sWzZv9i1squxbNm/2LQBsadwdDAAAAKAGhEAAAAAANdCyawIVRTE6yflJhidZVhTFuCTvqqrqqVa1CQAAAMCatfLC0HemeWEkAAAAALqZ6WAAAAAANSAEAgAAAKgBIRAAAABADQiBAAAAAGpACAQAAABQA0IgAAAAgBoQAgEAAADUgBAIAAAAoAaEQAAAAAA1IAQCAAAAqAEhEAAAAEANCIEAAAAAakAIBAAAAFADQiAAAACAGhACAQAAANSAEAgAAACgBoRAAAAAADUgBAIAAACoASEQAAAAQA0IgQAAAABqQAgEAAAAUANCIAAAAIAaEAIBAAAA1IAQCAAAAKAGhEAAAAAANSAEAgAAAKgBIRAAAABADQiBAAAAAGpACAQAAABQA0IgAAAAgBoQAgEAAADUgBAIAAAAoAaEQAAAAAA1IAQCAAAAqAEhEAAAAEANCIEAAAAAakAIBAAAAFADQiAAAACAGhACAQAAANSAEAgAAACgBoRAAAAAADUgBAIAAACoASEQAAAAQA0IgQAAAABqQAgEAAAAUANCIAAAAIAaEAIBAAAA1EB7dxcAAGxaHvvcfp3dXUOr7HrmPW3dXQMAQHcxEggAAACgBowEAoAXafSp07bYkTJJcvWA7q4AAIBWMBIIAAAAoAaEQAAAAAA1IAQCAAAAqAEhEAAAAEANCIEAAAAAakAIBAAAAFADQiAAAACAGhACAQAAANSAEAgAAACgBoRAAAAAADUgBAIAAACoASEQAAAAQA20d3cBwEv32Of26+zuGlpl1zPvaevuGgAAALYkRgIBAAAA1IAQCAAAAKAGWjodrCiKf0ny2iSdST5RVdVvWtkerG70qdO22OlSSXL1gO6uAAAAgM1Fy0YCFUXxpiR7VlV1cJIPJvlKq9oCAAAAYN3aOjtbM1CiKIrPJXmsqqpvdX1/f5IDq6pakCRtbW1b9AgNAAAAgI2ts7NzrTfZaeV0sB2S3LnK93O6HlvwQkUBAAAAsGFtzAtDC30AAAAAukkrQ6AZaY78WWlYkj+3sD0AAAAA1qKVIdBPk4xLkqIo9k8yo6qqhS1sDwAAAIC1aFkIVFXVrUnuLIri1jTvDPbxVrUFrL+iKEzNBLpFURQbcxo6m7CiKHYuiuKtXV/bLviLoii2Lopix+6uA2BL1coLQ6eqqs+0cv11VBRFzyQHJPl1VVXusMYLKoqiR1VVHSu/t93QakVR/F2SPyaZbXujKIpDkgxNctmq+yLqpyvs6VFV1fIkr0zy7SQ+7JOiKNqqqursOs49I8mCoijOs89gXVbuU5J8I8lPklzhuIP1sfLkQ133Mc68bAZWPUNWVdWKJO9P8vqiKHp3X1VsalaO8CmKon3V0T6r7ty6zq6NL4rihO6okS1TURRtRVH0XGVfdXaS13Yd0A8oimJwd9bHxlUURY+u7WHlfuiVSUYl2aEoigOKojikKIpe3VgiG0nXvuGv3o+6AqBUVXVDko6iKPrV9SC8zoqi2GXV71d+cO86zp2eZOs0w2P4i9VHs6+yT7koySuSfLrreT7j8ldW3ya6tp3avve0dXYKSzclXRtoW9eb4OrLdkiyKMm/JTkiyfiqqq7ayCWyiSmKYvskxyd5oqqqy1Z5fJskS5PsmuQrVVW9rSiKfknOTfJvVVXd1i0Fs0VYfYTZass+k2SXJHOSHJfkZ0n+qaqq6RuxRDaSroPytnVsDwckeU+SfkmGJ1mQ5Oaqqr660Ypko1jXtlAUxYAkb0ny5iQ/qKrq1qIofpXkS1VVXblyJMhGLpmNaOX7RlEUOyd5V5JLq6p6smvZ6CS7p/l+sWeSCWmOHvxv2wZrUhRFe5JDkuyf5D+SzEtyY5JXV1W1uDtro3t1vRf1XHnSYS3POSrJ25LsneRTSW6vUyjU0ulgvLCu0KdzlTMgHastb0/zQ9Rnk9yZ5L+TfD3JzlVVXbWuD2JsmbqGSnesckC0IslWSUZ2LX93kvclGZDmNnNKkhVFURzZtc3snuTZjV44m7U1TCtcdYTZdklOSDIwySVJfpTkm0m+lORVae6/3pHkyxuzZlqj6+Cqx8qTFV37os6uZb2SbJvkn5LsluS/knwnSa8ky6qqemvXdMHJSYRAW4BV9w2rbQs9kuyV5A1JfpXmzUJ2TvK7JOOLong2yeVphgFXJmlb+Vq2TKtsJ38qiuKiJIO7Tk59Mc27CE9PMjbN64h2pBka/7cAqD66Rg7vUVXVbasfd3Qd/x7ctfy7ab7HXJzknCQPV1X1TFEUdyd5e1EUV68rAGDLs8qU0h5JxiT5bZKnu5a9LMn/SfM49Yo0T5IfnuTfk3whzeOWPmkOtqgFIVA3KIri8DQPoK9fbefWluYG+eY0P9iflebv6D1JXp2kZ5IPVVX146IoLiyKYoeqqmZu/B6wMa0+OmzVUWJFUYyoqurhoij+mOQNRVFslebIn89UVfXboigeS/PgelKSSUVRPJXkhjRHafxuY/eFzddq+6qtkhyWZL80z7r9XZrD9m9NUiZ5d5Il6borZNf2uUdRFAPcJXLz1/WBbNX90IA0w+alSZ5KMjPNA6/P5X9HAD2RpGdRFFsneTDNaUD7VFV130Yun79BURRD07wu4S9W/i2vtm8YmuTINE80vDLJQWkecC9KcnuSXyYZneY+449pBsaf7Hq5D/pbgNVD4qIoelZVtaIoij5pvm8MTHJVkmPSHMVxbJKHknwsyd8nuTTJmUnmJhlWFMVWVVU9t/F7wsbSdXmLEVVV3Z/mNjI+yTu6Ro29rKqqeUVR9E1yWZIZSRYWRTEpyQ+6vr+3qqpnulb3izSnH/9nmu9HbKGKoujfFfy1VVX1VwMqiqL4+yRvLYrisiR/TnJe1/9tac6euC/NUclHdj32YJJa7WeEQC1WFMUrkrw2zZ3azDTf+O5K8wPSymHyO6eZQP5Xmju+76d50Pz1NIen9U3Su6qqp9M8W5Ikj3at9xqjgbYsXR+o3l9V1VeSNY4OG5tkhyRvTbJfURSfTzI/zQ9geyX5aZKPFEUxsusl76mq6sNFUfwyzQ/oP0nym43SGTYra7tIXtfB1xvSPEj7epI9kvxrkpPTHG12WZL9ut6Mf1xV1bNFUfw5zUAyae6vXp5kRJr7PzZxK6+7sOoZ+FXOsr0izRMW+yc5tytwPjFJWVXVRUVR/DDJhVVVPZZkStdrR6UZCuxQVdUjRVE8mOSNaR6IsQlbbVuYkqR3micT0jWt581pjvL6fppn6S9IMyC+I80TWD+tquqxVaYj/zLN/cdBVVVdUBRFR1EUO1dV9aeN3DX+RkVR7JTmh6g3pPmB6pKqqm5OV0hcFMXIqqoe6hoh+PU0RwTOTnOEz8/S3Hbak9yb5rHLb9IMhF6ZpEpzH7NyhBBbgK5j3DemGdR8o6qquVVVLS2K4tSiKD6U5jHCoUVRTElz5Mbsoig+meTANI9fv53mZ6Mjk/w6yTVpbkc/62rit/nfkWRsYYqi+Ic0jzvHJJlbFMXPqqpa1rXsVWnuY+5Psn2So9I80fDGNE8ynJbkn9MMoCel+V51eJrHpick+Xma17SsBSFQC3V9WL82zakQ30yyLM3pERem+WH91DSHof0gyflpng27L803vLFpBkEj0jyQOjjJDUVR7Jfm7+2yJMcWRfFYVVX/sxG7xQZWFMWBaZ4NOyDNofG/TPJfq1zAbEz+d2TFaWkGiq9J8s4kg9NMtOckmZXmm+ofkgzumnYxOs1tMF3/7931/Ckt7xibtKIo9kjzjfGpqqquSZ53Rr9vVVWLi6IYnuZ0nnuS9C+K4qQ0g+pZSR7sOsN7T5qh5NVVVc3rWsXP0zxo+7c0Q6Ct0/xAKATaxBRF0T/Ng+03pTkc+gtVVT2+yvLhaYY3/9310P+X5qiOS5KcXjQvNP+TNM+mJc3rMrwqyc+6pjQPTvMDYvL/t3ffUXqXZRrHvwkxVOkgrAUVlntBkEUCuGHpIEtHRaSjsIg0A6yKdFFYpYt0ZIMgTUVCUSmBIB1CkdDkQqQGwxKKEqQGsn9czwvjJEHYk5kkM9fnnJyQmbee98f7e373cxef0x7Du27r4okuMQPpZF5U1SBJEzuBwBYMGos/zyXbv0/F/6+/WlUHAMcBr0m6v91nIjBXe+gdgGsk/aiqvgJ8sqr+BQcFNgd+lN4vM492Mb8/zvIajnfRd6mqm/D/1x/HGX8nAA/hNe1WkiaWe1wOxsGi5XEQ6beSjin3M9wAr3PnxBsICQLNxNomwH14LTsMB/juwRvZR+JgzlBgFF6fvgLMJ2mlqjoO+E8c5DkLXw89CewuaUw7Dod1ebq7cLZhso5ncq3873P4+nkV/D2zBTBE0kEtu/DNqloRX0f/Fa8v7sUl5wtKurGq1sXXTINxkPDLkp6sqgUlHd+e6wrgi1U1t6QX6QcSBOpZY3At8xGdH7Qo92E4yn0EXjg/L+mpqnoZp8WeBvxA0ui2gB4C7Ih33j6Ho9tn40X7zlW1ezKBZj7ts70Vf57D8eJ5b3wifBD39HkcBwePBubHQZwRwAqSxrVj5l58jNwFLIUzgl5vT7NEe64T8WJtOK6hXgSnRUY/0+qiz8HlpQ8Dn6gqJF1cVZ/Gi6nl8AX8iTh489/4mPwObux6NT4O18HH8C24r8eIthOzEw52f73cL+hJvBvzp157o/F+nIDXA1cDCwC7twzDOYCT8HfRs63U5zZ8/CwIbIh37FfCZYEbAD/GAecdq+oMfPwc0B5nQ7zJAXAuLhOK6axLxvI6+PMZWVXHt1LOgfiC6lkcrNm63e0ivHlxFC712xlYSdLhVfVIVa0q6QbgEVwSdhsODi5ZVfvigMGD+OLuELxwJwGgmcqcwJqSlu7ys1FVtR7eJNgJfzdsjzeu7gSWAe6W9HRVzYPXL6vjz3/lFix4BQeQ9sYX/Q/1yruJaa5LUPcQfN2yGrCHpNva78fiY2NZvAE6SNLIqloHn2PA3zXH4munZyRt0+67WFWtho+hf66qhSU9gzdMH8SZyAkezmTaOWd7vLYYCXwDB3iObFnm5wJ7VtX8OCvsaLzJdBxwBc5Q/7Kk08sTkVfFJYM3Aju1zc012npmcFUdhBMz5m7P0S8CQJAR8T1tLPB8VW0IbzdPvBNn+wzAF/tX4QUWOBL+iKQTWwBoI5ziOBx4uKpG4y/LX7T0yc0l7ZoA0MypNay7CadPnyrpZnwB/ji+gN4EN3xepv09DvfxuRWPWv6gpE4Z2IT29+x4oT6oqu7DvX92BX7Vvtg2wguwcZXxmX1edRul2myEJzOtL2lPYA9cOw/+vrlZ0hBgUWC/VqO/BP4eegUfh8viuvuV2/1+jht8HoMDRne3Xi8HARMkvSbpcklZzE9HbeGzcQtAd362At6N31HSWZKOlfQdSX/DO2d3SVofX6TthgNCb+A068twduvGuPfYMu1hr2y/+znOdJ3QjodDJZ0DIOlZSX/u8Tcd76plLN+LL7hOB76MA0E7V9V/AL8D9sQXYAvjz3o03rGfHweB9sCbEJu2hx3V5b9H44t8cDnQGLyZcSOwi6RvSrpSknruXUYPeQkYV+5R+b2q2rk8eGJbXB44CWcmj8EbVK/hDYZO2cZ+eD08FH9PLNZ+dhWws6RNJV0k6Y+9/cZi2mjlw5/D3w8DcGbGbS2DA3yeeAZnaByNg38A1+ENBFow+SP4+LmjlY1tibMQB0l6CgcJOue1zYAxkh6dyhooZjBVtVlVHdr+ORueDng3DgB/CH/+H62qm/H6ovDxMB6flx7Gm+FX4OuewVX1sfYYS7bM1BeAo6pqR1yJs2i77voF8D1Jq0jqVE70C8kE6kGtMdXteLfsN/B2hPMPeGd8E7yzfnC7y0hgaItyvoEP8q3abtwh+IIsu2R9y3V4t+yEcvPEP1fVdbiZ3cqS9mmf/Xo4yHMmLg/7I7BXVR2GUyQ7wcXl8c7Ht/Ai7KnOE7VF15dwT4bJ+r7EzK9TvtECzl0nyHWd4LMkMH9VdZrNj2vfMR/EC7AhbdG2CHBqO5FuBmwt6dmqurQ9xvnAp8qN+R5p5R3rA+d2SlQl/abb60upRy+pqoVx8OWVznGBg3cfxiXGnUzAWdufQ6vqEdy77qEWsHsTOKDtto7HtfLjcb+FS+UhBfvji7iPAx+ud5o9n1FVt+KSwdcA2i5tzFimlLF8C15/bAd8Wx7RfTvO2jkDp9yvhfu3DJS0S7vf0KpaEGeCnYP7dlyDF+9ztE2L00gJYJ8g94A7APfSWAiPd/8qDgbfh9e2LwEv403RP+B1ywB8TjkCZ4d1yojX6vU3Eb1hIs4AOh1PYYLWF7WtPR7FQcJXgVmr6p9wYHpQVX1a0j34vLSVpI1bsGA9nFl0a3ucU7o83xu0DKCsN2Z85WERWwLrV9UdeBNhZbyumIQ3KP4dV8MMxBmIj+NMofvxenQszuT5miS1deoK+PyzTys/3R9X06yMs4Z+ByDppl55ozOgBIF63o3AgV3+PQk3uvshHuP9PDBHVQ2RdAewX1VthyOWN7RFE8pkhL7qdvyZz6V3Jhv8BQd1PlVVy0s6q3PjtnhaAUe7t8Zfeh8FhrcL78n6Q3W5+B8jafXuv4+ZW+fzbeWllwE3dgJ8VbU08IqkR3HA5y18Ebcb7uNxA9Dpz3AAXrDfAxzWApIfkxu6LgKsUW7ueSOwC158HYp38JAnFZ75bq81C7KeVX8/JGAH/D1xbwsMzoePgUE4EDQOQNLNLTPocJzx9QawSlX9Fpds3Sppg/b4n8JNF5/FGxYr42zGP+Hz2ha0sp722Pf17DuOaeBJ3FxzNUnXt58tgXdZ3wIuqKq7cangqXhi04N4gX4ynr5yMM4SXAsfd/cBh7eg76O4FCT6ILlP2K3tO2QueYrTU8DiVbUM/j5YBbikBY0XwhdtPwNGyxPE0hy+b3ux/XkDeLmqPir3Yxks6XW8jn2m/XkSb3Reic8z6+I1yVG8Ezia4vdJl82vy3r4/cQ0JOnVct/BH+O15XN4TbEgziZeDrhH0jNV9RDOLP0tLj+/HCdULIUziBYtDysYhzPLvok3JAYAT0k6A6+BAxgwaVLW5D2p3A/jUmD1thCfF/fS2B0voL6Ba/FHS8rI7n6mZYZdCpzSyZqoqrVwf5VBeAF1DU7HfwmnSx+KT6ZnSFphCo+ZbIs+rh03k9RtchMwu1zvPARfsI3HJ8MDu5betFTsSTjldhEcWCwc2BkK7C3p+ar6KbAvPsHuhi/0DgKea+VC3V/XZBOloufVFCZEts/i4y0lfhjedLgJn2++K+nXU3icuXBPoBfbbXfAu/XDcVnP4cDxcqPFYXjRdYak53rszUWPq6rv4izSG3EZ10I4Q3VF3BB833a7zXCWz7p4cb0+Hvd9ML6AG9WCijkH9SNdsjU656bz8LG0DM4+vAv3uUz5Zz/UNqP2xt8TRwN3Sjq2y+9/CLwp6YCWUXwU3ii/Cnh1SuWANZVJpjFzqXcmjo7CJcfb4D5yt+Jm8Wfi9enncZbhV/HQieNxP6CN8Vr3ZzhD6DDcquC7wJwpM353CQL1gqrq1Bg+jQ/qk+WxqIu03fPox1q518DO7kZV7YVLK0YBwyStXZ6+8wyerPLXdru/4ZHd/9tKyd6cPu8gelp5QsKkKVzoz4snZfweZ2iMlrRoEPbx1QAACRJJREFUudHdHyRdWFUbALdLGv8uj78intrxY9wTZG3cF+Rc4EeSXpnCfTon71zw9ZKWibMy7us0AbhA0uXtdwNwgOeTks5tn+kFkhavqjF4I+Ivrdx4NHB6KxXrfI4LdY6RqhqMNyl+CMyLA0ifxo2cz5X0fG++7+hZVbU2br76ddyDYS+czXMh3rTaBm9AbAt8DZgHH4Pn651pgNEPVVXhC67n8XGxBD4ujq+q+SS9MD1fX0x/bZ1yIG5/8BgOBgm4GQ8LWBjYtZWbzwLMIWmyyV5Z5/ZN5QmRw/AQnIXxRuMLeI3T6YN6Ll6HzIv7qH6mql7Czb8700l/T7fN0Xh3CQL1ghbZXhE3vrtV0hvT9xXFjKQ8BeFUfNG1Ht5d34a2qJI0rtvtO31fVsEX/Tme+pip7XJ1uWD/LK5tXh6PX54g6QtVdQ/enT8NB2+unsrj74P7N5yGp7dsClwl6aSWJfQhSY9N4X5TDEZFzys38L0EN069Ay+SLsa77QvgHbNH8PfHtTjD8Dy8cDodHw9XVtVO+ELtFElPtMfeCJeXPoAXWUNwOfJB7feTZRpF39Eyli8G1pb0egsCXoX7FL6FR3TPg6fIXZNjIbpq5Rcb4BKO67Mmia7aBsVWwJaSNmnl5TvgEuIr8dpjbPf7ZJOpfyg3k78Gl3gtiwcODAZ+hXtWnlDulXoMPi+dhzNVC29wJjD4/5QgUMR01hbgZ+Mvwas7adVTuN3bF2I5MfYt7/Z5VtXcwI64DvolvEu/Ct5N20HS/VV1PS7b2g6n4S8ODJZ0ULug2whPeXqsPebHcLBoC7wbdy1evE/s9tyTlZ3F9NE+swskDe3ys5E4/Xk87vNzO95FWwz3eNoFN5GfBVhc0t5VtSbwfWCv1ofu7dKxdvt7gGu7B5+jb6uqEcAVkk5r/14aH0d3vFsWYUTEe9EyUneVJzJFAG+vce/CvXtGSBpTVafjzNORkjatqm2BB9SGjsS0kSBQxAwoQZ6+q6o+IOmNqfXPabuqC+IU12PxrsgGwIn4wv5p3PB3F+DXkn5dnpYxK27euhO+yD8cj+9eAPi8pO164e1FD2nHy6XAUZ0GvlV1Hi7P+k1V7Q5sjtPsZ8ENE9/EtfQ/wFlEX8TBv41xJtAFvf5GYoZUVdsDr+eYiIie0LLeN8PDCk5rmcUkk6N/q6rl8Hp3WGsU3tn0WgtvQmTARA9JEChiBtGma7yZ4E/fU1XL4ovxVYE/4+kEd7SeLLPjZsz34/r4bwE/wc3Aj8AX7+vgJs6di/zDgTWBiZKOraqVcD+f1YEn8MS4f8VTeZ4Hfinp0qm8tpR4zSS6NfBdDy+m966qRXFJ6c5tgsYIPPntMhwMWhKXmG6Ns74uScPEiIjoTa2H1OwZhBMd7ZjYU9IeKT3vXQkCRUT0kJa9MQfOyBmH+4ItgidlXIQDP/vi0acn412ymyRdXFUHAqtKWq+qTsUTD+7AwZ7/whMTlgH2bz2iJgCLAifh6U+P9t47jd7QpYHvHrgfmLr87kE8dXI2fGzsDGyPa+zPzQSviIiIiACPoI6IiB7QGhvuAnxA0sHtxw9U1Wy4f8/D7We/knRTVe2Ax7ADnALs2Hr6bCfp6wBVtRgOIgk3z1u8/ffawCuSdmi36zRWzCjVvuNu3LPn/Bb4GwDM0no57Y+Pi1eBb+Nyr/E48yciIiIiAkgQKCKip70MjIV3+gHhUp0tgUlA1xHLL+IGveC+LuDR3OdV1WV4ZOZxuM/P3cDJkv7WUmhHdx6k6yjVBH/6DknPVdWzuO/TacCALs28R0i6aPq9uoiIiIiYGSQIFBHRswYAE6pqVkmvtQydCVU1F/4OfhCPugQHhzasqq+2+43F/YD2Br6Ex2FO1iSve6AnjRb7tBHA6/D3n3t6iUVERETEe5EgUEREz3oA2ASPW34ImB1nB/0V+Age0b5jVX1C0iVVNQnYB/cRGg6sJ+kl4MzOA7YyoIEJ9vQ/ks6e3q8hIiIiImZeA6f3C4iI6OPuwqO6dwOQ9HKbFvYaMApn+7zEO0H5e3AA6DVgIzwpjKoa2HWsfAJAERERERHxfmU6WERED6uqxYEjgWeBTwDzAmdLOnEKtx0AHIObPnfGeU/sfruIiIiIiIj3K0GgiIheUlWbAY9JunsKvxuQvi4REREREdGTEgSKiOgF3YM8baJXJndFRERERESvSRAoIqIXJeMnIiIiIiKmlwSBIiIiIiIiIiL6gUwHi4iIiIiIiIjoBxIEioiIiIiIiIjoBxIEioiIiIiIiIjoBxIEioiIiH6hqtaoqgt78fl+WlUb9dbzRURERPwjCQJFRERERERERPQDg6b3C4iIiIjoCVX1AeAsYDHgVWA4MFdVnQMsB/xS0veqah3g+8DrwAvAFsBQYA/gLWAp4EJJh1bV74CRwFrAgsDGkp6oqsOBVYFZgBMlnd/ldXwMOAd4E6+9tpX0eE+//4iIiIjukgkUERERfdUOwNOSVgF+AswNLA18Dfg3YM92u/mArSWtDrwIrNd+vhLwlW63BXhR0trA5cAXqmpVYDFJq+Hg0IFVNXuX228OjJS0JjAMWHRav9GIiIiI9yKZQBEREdFXfQa4BkDSBVW1BnCXpJcBqmpAu9144IyqGgR8EhgFTOh2266Pe0P7eyywAM4a+mzLEgJvsnUN9FwFjKiqeXFG0S3T8D1GREREvGfJBIqIiIi+6k0mX+tMnMLthgN7tEygS/7Bbbv/fAAuI/sfSWu0P0tJeqRzA0n34fKzG4AfVNX27/N9REREREwTCQJFREREX3U7Ls+iTekaOpXbzQM80TJ11gQGv8/nuQ3YuKoGVtVsVXVC119W1ZbAMpIuBg4EhrzPx4+IiIiYJhIEioiIiL7qAmDOqroO2AuYWjPmk4CbgNOBI4H9eB99eyTdDFwL3AJcD9zZ7SYPASdW1SjgEOCU9/EeIiIiIqaZAZMmTZreryEiIiIiIiIiInpYMoEiIiIiIiIiIvqBBIEiIiIiIiIiIvqBBIEiIiIiIiIiIvqBBIEiIiIiIiIiIvqBBIEiIiIiIiIiIvqBBIEiIiIiIiIiIvqBBIEiIiIiIiIiIvqB/wPxyflyOwqP0gAAAABJRU5ErkJggg==\n","text/plain":["<Figure size 1440x504 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]}]}