{
  "metadata": {
    "version": 1,
    "disable_limits": false
  },
  "nodes": [
    {
      "node_id": "e405a68e-7b86-438c-9e98-9d2b1963e167",
      "type": "SOURCE",
      "operator": "sagemaker.s3_source_0.1",
      "parameters": {
        "dataset_definition": {
          "__typename": "S3CreateDatasetDefinitionOutput",
          "datasetSourceType": "S3",
          "name": "signup_attempts.csv",
          "description": null,
          "s3ExecutionContext": {
            "__typename": "S3ExecutionContext",
            "s3Uri": "s3://${bucket}/${prefix}/data/raw/signup_attempts.csv",
            "s3ContentType": "csv",
            "s3HasHeader": true
          }
        }
      },
      "inputs": [],
      "outputs": [
        {
          "name": "default",
          "sampling": {
            "sampling_method": "sample_by_limit",
            "limit_rows": 50000
          }
        }
      ]
    },
    {
      "node_id": "91372460-ddbc-4dd7-be64-db0a12c30dad",
      "type": "TRANSFORM",
      "operator": "sagemaker.spark.infer_and_cast_type_0.1",
      "parameters": {},
      "trained_parameters": {
        "schema": {
          "ip_address": "string",
          "email_address": "string",
          "user_agent": "string",
          "customer_city": "string",
          "customer_state": "string",
          "customer_postal": "long",
          "customer_name": "string",
          "customer_address": "string",
          "phone_number": "string",
          "timestamp": "long"
        }
      },
      "inputs": [
        {
          "name": "default",
          "node_id": "e405a68e-7b86-438c-9e98-9d2b1963e167",
          "output_name": "default"
        }
      ],
      "outputs": [
        {
          "name": "default"
        }
      ]
    },
    {
      "node_id": "dbe6a4eb-f697-4e8e-b7ef-9faef77e1020",
      "type": "TRANSFORM",
      "operator": "sagemaker.spark.custom_pandas_0.1",
      "parameters": {
        "code": "# Table is available as variable `df`\nimport pandas as pd\ndf['timestamp'] = pd.to_datetime(df['timestamp'], unit='ms').dt.strftime('%m/%d/%y %H:%M:%S')"
      },
      "inputs": [
        {
          "name": "df",
          "node_id": "91372460-ddbc-4dd7-be64-db0a12c30dad",
          "output_name": "default"
        }
      ],
      "outputs": [
        {
          "name": "default"
        }
      ]
    },
    {
      "node_id": "572880dd-87da-4b63-888d-0f24cc4eb158",
      "type": "TRANSFORM",
      "operator": "sagemaker.spark.manage_columns_0.1",
      "parameters": {
        "operator": "Rename column",
        "rename_column_parameters": {
          "input_column": "timestamp",
          "new_name": "EVENT_TIMESTAMP"
        },
        "drop_column_parameters": {}
      },
      "inputs": [
        {
          "name": "df",
          "node_id": "dbe6a4eb-f697-4e8e-b7ef-9faef77e1020",
          "output_name": "default"
        }
      ],
      "outputs": [
        {
          "name": "default"
        }
      ]
    },
    {
      "node_id": "47d9c8e0-9478-416f-9ef8-604b4ca7eada",
      "type": "TRANSFORM",
      "operator": "sagemaker.spark.handle_missing_0.1",
      "parameters": {
        "operator": "Fill missing",
        "fill_missing_parameters": {
          "input_column": "email_address",
          "fill_value": "synth_missing@email.com"
        },
        "impute_parameters": {
          "column_type": "Numeric",
          "numeric_parameters": {
            "strategy": "Approximate Median"
          }
        }
      },
      "inputs": [
        {
          "name": "df",
          "node_id": "572880dd-87da-4b63-888d-0f24cc4eb158",
          "output_name": "default"
        }
      ],
      "outputs": [
        {
          "name": "default"
        }
      ]
    },
    {
      "node_id": "7f2cefd2-5538-4f09-9f1e-a1d16fbe8b39",
      "type": "TRANSFORM",
      "operator": "sagemaker.spark.handle_missing_0.1",
      "parameters": {
        "operator": "Fill missing",
        "fill_missing_parameters": {
          "input_column": "user_agent",
          "fill_value": "Missing"
        },
        "impute_parameters": {
          "column_type": "Numeric",
          "numeric_parameters": {
            "strategy": "Approximate Median"
          }
        }
      },
      "inputs": [
        {
          "name": "df",
          "node_id": "47d9c8e0-9478-416f-9ef8-604b4ca7eada",
          "output_name": "default"
        }
      ],
      "outputs": [
        {
          "name": "default"
        }
      ]
    },
    {
      "node_id": "5481d4b8-03fc-48c9-8051-bdc5333f2098",
      "type": "TRANSFORM",
      "operator": "sagemaker.spark.handle_missing_0.1",
      "parameters": {
        "operator": "Fill missing",
        "fill_missing_parameters": {
          "input_column": "customer_city",
          "fill_value": "Missing"
        },
        "impute_parameters": {
          "column_type": "Numeric",
          "numeric_parameters": {
            "strategy": "Approximate Median"
          }
        }
      },
      "inputs": [
        {
          "name": "df",
          "node_id": "7f2cefd2-5538-4f09-9f1e-a1d16fbe8b39",
          "output_name": "default"
        }
      ],
      "outputs": [
        {
          "name": "default"
        }
      ]
    },
    {
      "node_id": "bab72342-c6c1-48df-a259-751c0a7df470",
      "type": "TRANSFORM",
      "operator": "sagemaker.spark.handle_missing_0.1",
      "parameters": {
        "operator": "Fill missing",
        "fill_missing_parameters": {
          "input_column": "customer_state",
          "fill_value": "Missing"
        },
        "impute_parameters": {
          "column_type": "Numeric",
          "numeric_parameters": {
            "strategy": "Approximate Median"
          }
        }
      },
      "inputs": [
        {
          "name": "df",
          "node_id": "5481d4b8-03fc-48c9-8051-bdc5333f2098",
          "output_name": "default"
        }
      ],
      "outputs": [
        {
          "name": "default"
        }
      ]
    },
    {
      "node_id": "a6ddb7dd-2817-4030-b960-b072b1cd68f9",
      "type": "TRANSFORM",
      "operator": "sagemaker.spark.handle_missing_0.1",
      "parameters": {
        "operator": "Fill missing",
        "fill_missing_parameters": {
          "input_column": "customer_postal",
          "fill_value": "0"
        },
        "impute_parameters": {
          "column_type": "Numeric",
          "numeric_parameters": {
            "strategy": "Approximate Median"
          }
        }
      },
      "inputs": [
        {
          "name": "df",
          "node_id": "bab72342-c6c1-48df-a259-751c0a7df470",
          "output_name": "default"
        }
      ],
      "outputs": [
        {
          "name": "default"
        }
      ]
    },
    {
      "node_id": "e076b7cc-d5f2-4996-852e-59d601b19a9d",
      "type": "TRANSFORM",
      "operator": "sagemaker.spark.handle_missing_0.1",
      "parameters": {
        "operator": "Fill missing",
        "fill_missing_parameters": {
          "input_column": "customer_name",
          "fill_value": "Missing"
        },
        "impute_parameters": {
          "column_type": "Numeric",
          "numeric_parameters": {
            "strategy": "Approximate Median"
          }
        }
      },
      "inputs": [
        {
          "name": "df",
          "node_id": "a6ddb7dd-2817-4030-b960-b072b1cd68f9",
          "output_name": "default"
        }
      ],
      "outputs": [
        {
          "name": "default"
        }
      ]
    },
    {
      "node_id": "02f86fd3-c64b-4dbb-9736-3e9e0ba1c0ce",
      "type": "TRANSFORM",
      "operator": "sagemaker.spark.handle_missing_0.1",
      "parameters": {
        "operator": "Fill missing",
        "fill_missing_parameters": {
          "input_column": "customer_address",
          "fill_value": "Missing"
        },
        "impute_parameters": {
          "column_type": "Numeric",
          "numeric_parameters": {
            "strategy": "Approximate Median"
          }
        }
      },
      "inputs": [
        {
          "name": "df",
          "node_id": "e076b7cc-d5f2-4996-852e-59d601b19a9d",
          "output_name": "default"
        }
      ],
      "outputs": [
        {
          "name": "default"
        }
      ]
    },
    {
      "node_id": "13a61851-f189-4b0d-8aeb-64dd40965e60",
      "type": "TRANSFORM",
      "operator": "sagemaker.spark.handle_missing_0.1",
      "parameters": {
        "operator": "Fill missing",
        "fill_missing_parameters": {
          "input_column": "phone_number",
          "fill_value": "Missing"
        },
        "impute_parameters": {
          "column_type": "Numeric",
          "numeric_parameters": {
            "strategy": "Approximate Median"
          }
        }
      },
      "inputs": [
        {
          "name": "df",
          "node_id": "02f86fd3-c64b-4dbb-9736-3e9e0ba1c0ce",
          "output_name": "default"
        }
      ],
      "outputs": [
        {
          "name": "default"
        }
      ]
    },
    {
      "node_id": "ec09720f-c544-4a30-b0c1-9e4730889f6b",
      "type": "TRANSFORM",
      "operator": "sagemaker.spark.handle_missing_0.1",
      "parameters": {
        "operator": "Fill missing",
        "fill_missing_parameters": {
          "input_column": "ip_address",
          "fill_value": "0.0.0.0"
        },
        "impute_parameters": {
          "column_type": "Numeric",
          "numeric_parameters": {
            "strategy": "Approximate Median"
          }
        }
      },
      "inputs": [
        {
          "name": "df",
          "node_id": "13a61851-f189-4b0d-8aeb-64dd40965e60",
          "output_name": "default"
        }
      ],
      "outputs": [
        {
          "name": "default"
        }
      ]
    },
    {
      "node_id": "b70a84b8-6570-487f-9e3c-b72f68886a28",
      "type": "TRANSFORM",
      "operator": "sagemaker.spark.custom_pandas_0.1",
      "parameters": {
        "code": "# Table is available as variable `df`\nimport pandas as pd\ndf['EventTime'] = pd.to_datetime('now').timestamp()"
      },
      "inputs": [
        {
          "name": "df",
          "node_id": "ec09720f-c544-4a30-b0c1-9e4730889f6b",
          "output_name": "default"
        }
      ],
      "outputs": [
        {
          "name": "default"
        }
      ]
    }
  ]
}