Show Menu
TOPICS×

Insights

Insights contain metrics which are used to empower a data scientist to evaluate and choose optimal ML models by displaying relevant evaluation metrics.

Retrieve a list of Insights

You can retrieve a list of Insights by performing a single GET request to the insights endpoint. To help filter results, you can specify query parameters in the request path. For a list of available queries, refer to the appendix section on query parameters for asset retrieval .
API Format
GET /insights

Request
curl -X GET \
  https://platform.adobe.io/data/sensei/insights \
    -H 'Authorization: Bearer {ACCESS_TOKEN}' \
    -H 'x-api-key: {API_KEY}' \
    -H 'x-gw-ims-org-id: {IMS_ORG}' \
    -H 'x-sandbox-name: {SANDBOX_NAME}'

Response
A successful response returns a payload that includes a list of insights and each insight has unique identifier ( id ). Additionally, you will receive context which contains the unique identifiers that are associated with that particular insight following with the Insights events and metrics data.
{
    "children": [
        {
            "id": "08b8d174-6b0d-4d7e-acd8-1c4c908e14b2",
            "context": {
                "experimentId": "5cb25a2d-2cbd-4c99-a619-8ddae5250a7b",
                "experimentRunId": "33408593-2871-4198-a812-6d1b7d939cda",
                "modelId": "15c53796-bd6b-4e09-b51d-7296aa20af71"
            },
            "events": {
                "name": "fit",
                "eventValues": {
                    "algorithm": null,
                    "ratio": "0.8"
                }
            },
            "metrics": [
                {
                    "name": "MAPE",
                    "value": "0.0111111111111",
                    "valueType": "double"
                }
            ],
            "created": "2019-01-01T00:00:00.000Z",
            "updated": "2019-01-02T00:00:00.000Z"
        },
        {
            "id": "08b8d174-6b0d-4d7e-acd8-1c4c908e14b2",
            "context": {
                "experimentId": "5cb25a2d-2cbd-4c99-a619-8ddae5250a7b",
                "experimentRunId": "33408593-2871-4198-a812-6d1b7d939cda",
                "modelId": "15c53796-bd6b-4e09-b51d-7296aa20af71"
            },
            "events": {
                "name": "fit",
                "eventValues": {
                    "algorithm": null,
                    "ratio": "0.8"
                }
            },
            "metrics": [
                {
                    "name": "MAPE",
                    "value": "0.0111111111111",
                    "valueType": "double"
                }
            ],
            "created": "2019-01-01T00:00:00.000Z",
            "updated": "2019-01-02T00:00:00.000Z"
            }
        ],
    "_page": {
        "count": 2
    }
}

Property
Description
id
The ID corresponding to the Insight.
experimentId
A valid Experiment ID.
experimentRunId
A valid Experiment Run ID.
modelId
A valid Model ID.

Retrieve a specific Insight

To look up a particular insight make a GET request and provide a valid {INSIGHT_ID} in the request path. To help filter results, you can specify query parameters in the request path. For a list of available queries, refer to the appendix section on query parameters for asset retrieval .
API Format
GET /insights/{INSIGHT_ID}

Parameter
Description
{INSIGHT_ID}
The unique identifier of a Sensei insight.
Request
curl -X GET \
  https://platform.adobe.io/data/sensei/insights/08b8d174-6b0d-4d7e-acd8-1c4c908e14b2 \
    -H 'Authorization: Bearer {ACCESS_TOKEN}' \
    -H 'x-api-key: {API_KEY}' \
    -H 'x-gw-ims-org-id: {IMS_ORG}' \
    -H 'x-sandbox-name: {SANDBOX_NAME}'

Response
A successful response returns a payload that includes the insights unique identifier ( id ). Additionally you will receive context which contains the unique identifiers that are associated with the particular insight following with the Insights events and metrics data.
{
    "id": "08b8d174-6b0d-4d7e-acd8-1c4c908e14b2",
    "context": {
        "experimentId": "5cb25a2d-2cbd-4c99-a619-8ddae5250a7b",
        "experimentRunId": "33408593-2871-4198-a812-6d1b7d939cda",
        "modelId": "15c53796-bd6b-4e09-b51d-7296aa20af71"
    },
    "events": {
        "name": "fit",
        "eventValues": {
            "algorithm": null,
            "ratio": "0.8"
        }
    },
    "metrics": [
        {
            "name": "MAPE",
            "value": "0.0111111111111",
            "valueType": "double"
        }
    ],
    "created": "2019-01-01T00:00:00.000Z",
    "updated": "2019-01-02T00:00:00.000Z"
}

Property
Description
id
The ID corresponding to the Insight.
experimentId
A valid Experiment ID.
experimentRunId
A valid Experiment Run ID.
modelId
A valid Model ID.

Add a new Model insight

You can create a new Model insight by performing a POST request and a payload that provides context, events,and metrics for the new Model insight. The context field used to create a new Model insight is not required to have existing services attached to it but you can choose to create the new Model insight with existing services by providing one or more of the corresponding IDs:
"context": {
    "clientId": "f1ab3164-e688-433d-99ef-077b2be84731",
    "notebookId": "T4ab3164-e658-443d-97ef-022b2be84999",
    "experimentId": "5cb25a2d-2cbd-4c99-a619-8ddae5250a7b",
    "engineId": "22f4166f-85ba-4130-a995-a2b8e1edde32",
    "mlInstanceId": "46986c8f-7739-4376-8509-0178bdf32cda",
    "experimentRunId": "33408593-2871-4198-a812-6d1b7d939cda",
    "modelId": "15c53796-bd6b-4e09-b51d-7296aa20af71",
    "dataSetId": "5ee3cd7f2d34011913c56941"
  }

API Format
POST /insights

Request
curl -X POST \
  https://platform.adobe.io/data/sensei/insights \
    -H 'Authorization: Bearer {ACCESS_TOKEN}' \
    -H 'x-api-key: {API_KEY}' \
    -H 'x-gw-ims-org-id: {IMS_ORG}' \
    -H 'x-sandbox-name: {SANDBOX_NAME}'
    -H `Content-Type: application/vnd.adobe.platform.sensei+json;profile=mlInstance.v1.json`
    -d {
    "context": {
        "experimentId": "5cb25a2d-2cbd-4c99-a619-8ddae5250a7b",
        "experimentRunId": "33408593-2871-4198-a812-6d1b7d939cda",
        "modelId": "15c53796-bd6b-4e09-b51d-7296aa20af71"
    },
    "events": {
        "name": "fit2",
        "eventValues": {
            "algorithm": null,
            "ratio": "0.99"
        }
    },
    "metrics": [
        {
            "name": "MAPE2",
            "value": "0.11111111111",
            "valueType": "double"
        }
    ],
    "created": "2019-01-01T00:00:00.000Z",
    "updated": "2019-01-02T00:00:00.000Z"
}

Response
A successful response will return a payload that has an {INSIGHT_ID} and any parameters that you provided in the initial request.
{
    "id": "08b8d174-6b0d-4d7e-acd8-1c4c908e14b2",
    "context": {
        "experimentId": "5cb25a2d-2cbd-4c99-a619-8ddae5250a7b",
        "experimentRunId": "33408593-2871-4198-a812-6d1b7d939cda",
        "modelId": "15c53796-bd6b-4e09-b51d-7296aa20af71"
    },
    "events": {
        "name": "fit2",
        "eventValues": {
            "algorithm": null,
            "ratio": "0.99"
        }
    },
    "metrics": [
        {
            "name": "MAPE2",
            "value": "0.11111111111",
            "valueType": "double"
        }
    ],
    "created": "2019-01-01T00:00:00.000Z",
    "updated": "2019-01-02T00:00:00.000Z"
}

Property
Description
insightId
The unique ID that is created for this particular insight when a successful POST request is made.

Retrieve a list of default metrics for algorithms

You can retrieve a list of all your algorithm's and default metrics by performing a single GET request to the metrics endpoint. To query a particular metric make a GET request and provide a valid {ALGORITHM} in the request path.
API Format
GET /insights/metrics
GET /insights/metrics?algorithm={ALGORITHM}

Parameter
Description
{ALGORITHM}
The identifier of the algorithm type.
Request
The following request contains a query and retrieves a specific metric by using the algorithm identifier {ALGORITHM}
curl -X GET \
  'https://platform.adobe.io/data/sensei/insights/metrics?algorithm={ALGORITHM}' \
    -H 'Authorization: Bearer {ACCESS_TOKEN}' \
    -H 'x-api-key: {API_KEY}' \
    -H 'x-gw-ims-org-id: {IMS_ORG}' \
    -H 'x-sandbox-name: {SANDBOX_NAME}'


Response
A successful response returns a payload that includes the algorithm unique identifier and an array of default metrics.
{
    "children": [
        {
            "algorithm": "15c53796-bd6b-4e09-b51d-7296aa20af71",
            "defaultMetrics": [
                "f-score",
                "auroc",
                "roc",
                "precision",
                "recall",
                "accuracy",
                "confusion matrix"
            ]
        }
    ]
}