Analyze rodent infestation utilizing Amazon SageMaker geospatial capabilities

Rodents resembling rats and mice are related to a variety of well being dangers and are identified to unfold greater than 35 illnesses. Figuring out areas of excessive rodent exercise might help native authorities and pest management organizations plan for interventions successfully and exterminate the rodents.

On this submit, we present monitor and visualize a rodent inhabitants utilizing Amazon SageMaker geospatial capabilities. We then visualize rodent infestation results on vegetation and our bodies of water. Lastly, we correlate and visualize the variety of monkey pox circumstances reported with rodent sightings in a area. Amazon SageMaker makes it simpler for knowledge scientists and machine studying (ML) engineers to construct, prepare, and deploy fashions utilizing geospatial knowledge. The instrument makes it simpler to entry geospatial knowledge sources, run purpose-built processing operations, apply pre-trained ML fashions, and use built-in visualization instruments sooner and at scale.

Pocket book

First, we use an Amazon SageMaker Studio pocket book with a geospatial picture by following the steps outlined in Getting Started with Amazon SageMaker geospatial capabilities.

Knowledge entry

The geospatial picture comes preinstalled with SageMaker geospatial capabilities that make it simpler to complement knowledge for geospatial evaluation and ML. For our submit, we use satellite tv for pc pictures from Sentinel-2 and the rodent activity and monkeypox datasets from open-source NYC open data.

First, we use the rodent exercise and extract the latitude and longitude of rodent sightings and inspections. Then we enrich this location info with human-readable road addresses. We create a vector enrichment job (VEJ) within the SageMaker Studio pocket book to run a reverse geocoding operation so that you could convert geographic coordinates (latitude, longitude) to human-readable addresses, powered by Amazon Location Service. We create the VEJ as follows:

import boto3
import botocore
import sagemaker
import sagemaker_geospatial_map

area = boto3.Session().region_name
session = botocore.session.get_session()
execution_role = sagemaker.get_execution_role()

sg_client= session.create_client(
    service_name="sagemaker-geospatial",
    region_name=area
)
response = sg_client.start_vector_enrichment_job(
    ExecutionRoleArn=execution_role,
    InputConfig={
        'DataSourceConfig': {
            'S3Data': {
                'S3Uri': 's3://<bucket>/pattern/rodent.csv'
            }
        },
        'DocumentType': 'CSV'
    },
    JobConfig={
        "ReverseGeocodingConfig": { 
         "XAttributeName": "longitude",
         "YAttributeName": "latitude"
      }
    },
    Title="vej-reversegeo",
)

my_vej_arn = response['Arn']

Visualize rodent exercise in a area

Now we are able to use SageMaker geospatial capabilities to visualise rodent sightings. After the VEJ is full, we export the output of the job to an Amazon S3 bucket.

sg_client.export_vector_enrichment_job(
    Arn=my_vej_arn,
    ExecutionRoleArn=execution_role,
    OutputConfig={
        'S3Data': {
            'S3Uri': 's3://<bucket>/reversegeo/'
        }
    }
)

When the export is full, you will notice the output CSV file in your Amazon Simple Storage Service (Amazon S3) bucket, which consists of your enter knowledge (longitude and latitude coordinates) together with extra columns: handle quantity, nation, label, municipality, neighborhood, postal code, and area of that location appended on the finish.

From the output file generated by VEJ, we are able to use SageMaker geospatial capabilities to overlay the output on a base map and supply layered visualization to make collaboration simpler. SageMaker geospatial capabilities present built-in visualization tooling powered by Foursquare Studio, which natively works from inside a SageMaker pocket book by way of the SageMaker geospatial Map SDK. Beneath, we are able to visualize the rodent sightings and likewise get the human readable addresses for every of the info factors. The handle info of every of the rodent sightings knowledge factors will be helpful for rodent inspection and remedy functions.

Analyze the consequences of rodent infestation on vegetation and our bodies of water

To research the consequences of rodent infestation on vegetation and our bodies of water, we have to classify every location as vegetation, water, and naked floor. Let’s have a look at how we are able to use these geospatial capabilities to carry out this evaluation.

The brand new geospatial capabilities in SageMaker supply simpler entry to geospatial knowledge resembling Sentinel-2 and Landsat 8. Constructed-in geospatial dataset entry saves weeks of effort in any other case misplaced to amassing and processing knowledge from numerous knowledge suppliers and distributors. Additionally, these geospatial capabilities supply a pre-trained Land Use Land Cowl (LULC) segmentation mannequin to establish the bodily materials, resembling vegetation, water, and naked floor, on the earth floor.

We use this LULC ML mannequin to research the consequences of rodent inhabitants on vegetation and our bodies of water.

Within the following code snippet, we first outline the realm of curiosity coordinates (aoi_coords) of New York Metropolis. Then we create an Earth Commentary Job (EOJ) and choose the LULC operation. SageMaker downloads and preprocesses the satellite tv for pc picture knowledge for the EOJ. Subsequent, SageMaker robotically runs mannequin inference for the EOJ. The runtime of the EOJ will differ from a number of minutes to hours relying on the variety of pictures processed. You possibly can monitor the standing of EOJs utilizing the get_earth_observation_job operate, and visualize the enter and output of the EOJ within the map.

aoi_coords = [
    [
            [
              -74.13513011934334,
              40.87856296920188
            ],
            [
              -74.13513011934334,
              40.565792636343616
            ],
            [
              -73.8247144462764,
              40.565792636343616
            ],
            [
              -73.8247144462764,
              40.87856296920188
            ],
            [
              -74.13513011934334,
              40.87856296920188
            ]
    ]
]

eoj_input_config = {
    "RasterDataCollectionQuery": {
        "RasterDataCollectionArn": "arn:aws:sagemaker-geospatial:us-west-2:378778860802:raster-data-collection/public/nmqj48dcu3g7ayw8",
        "AreaOfInterest": {
            "AreaOfInterestGeometry": {
                "PolygonGeometry": {
                    "Coordinates": aoi_coords
                }
            }
        },
        "TimeRangeFilter": {
            "StartTime": "2023-01-01T00:00:00Z",
            "EndTime": "2023-02-28T23:59:59Z",
        },
        "PropertyFilters": {
            "Properties": [{"Property": {"EoCloudCover": {"LowerBound": 0, "UpperBound": 2.0}}}],
            "LogicalOperator": "AND",
        },
    }
}
eoj_config = {
  "LandCoverSegmentationConfig": {}
}

response = geospatial_client.start_earth_observation_job(
    Title="eoj-rodent-infestation-lulc-example",
    InputConfig=eoj_input_config,
    JobConfig=eoj_config,
    ExecutionRoleArn=execution_role,
)
eoj_arn = response["Arn"]
eoj_arn

Map = sagemaker_geospatial_map.create_map()
Map.set_sagemaker_geospatial_client(sg_client)

Map.render()

time_range_filter = {
    "start_date": "2023-01-01T00:00:00Z",
    "end_date": "2023-02-28T23:59:59Z",
}


config = {"preset": "singleBand", "band_name": "masks"}
output_layer = Map.visualize_eoj_output(
    Arn=eoj_arn, config=config, time_range_filter=time_range_filter
)

To visualise the rodent inhabitants with respect to vegetation, we overlay the rodent inhabitants and sighting knowledge on the land cowl segmentation mannequin predictions. This visualization might help us find the inhabitants of rodents and analyze it on vegetation and our bodies of water.

Visualize monkeypox circumstances and corelating with rodent knowledge

To visualise the relation between the monkeypox circumstances and rodent sightings, we add the monkeypox dataset and the geoJSON file for New York City borough boundaries. See the next code:

nybb = pd.read_csv("./nybb.csv")
monkeypox = pd.read_csv("./monkeypox.csv")
dataset = Map.add_dataset({
    "knowledge": nybb
}, auto_create_layers=False)
dataset = Map.add_dataset({
    "knowledge": monkeypox
}, auto_create_layers=False)

Inside a SageMaker Studio pocket book, we are able to use the visualization instrument powered by Foursquare so as to add layers within the map and add charts. Right here, we added the monkeypox knowledge as a chart to point out the variety of monkeypox circumstances for every of the boroughs. To see the correlation between monkeypox circumstances and rodent sightings, now we have added the borough boundaries as a polygon layer and added the heatmap layer that represents rodent exercise. The borough boundary layer is coloured to match the monkeypox knowledge chart. As we are able to see, the borough of Manhattan reveals a excessive focus of rodent sightings and data the best variety of monkeypox circumstances, adopted by Brooklyn.

That is supported by a easy statistical evaluation of calculating the correlation between the focus of rodent sightings and monkeypox circumstances in every borough. The calculation produced an r worth of 0.714, which suggests a optimistic correlation.

r = np.corrcoef(borough_stats['Concentration (sightings per square km)'], borough_stats['Monkeypox Cases'])

Conclusion

On this submit, we demonstrated how you should use SageMaker geospatial capabilities to get detailed addresses of rodent sightings and visualize the rodent results on vegetation and our bodies of water. This might help native authorities and pest management organizations plan for interventions successfully and exterminate rodents. We additionally correlated the rodent sightings to monkeypox circumstances within the space with the built-in visualization instrument. By using vector enrichment and EOJs together with the built-in visualization instruments, SageMaker geospatial capabilities remove the challenges of dealing with large-scale geospatial datasets, mannequin coaching, and inference, and supply the flexibility to quickly discover predictions and geospatial knowledge on an interactive map utilizing 3D accelerated graphics and built-in visualization instruments.

You may get began with SageMaker geospatial capabilities in two methods:

To study extra, go to Amazon SageMaker geospatial capabilities and Getting Started with Amazon SageMaker geospatial capabilitites. Additionally, go to our GitHub repo, which has a number of instance notebooks on SageMaker geospatial capabilities.

In regards to the authors

Bunny Kaushik is a Options Architect at AWS. He’s enthusiastic about constructing AI/ML options and serving to clients innovate on the AWS platform. Outdoors of labor, he enjoys climbing, mountaineering, and swimming.

Clarisse Vigal is a Sr. Technical Account Supervisor at AWS, targeted on serving to clients speed up their cloud adoption journey. Outdoors of labor, Clarisse enjoys touring, climbing, and studying sci-fi thrillers.

Veda Raman is a Senior Specialist Options Architect for machine studying primarily based in Maryland. Veda works with clients to assist them architect environment friendly, safe and scalable machine studying functions. Veda is considering serving to clients leverage serverless applied sciences for Machine studying.