Perform classification with a boosted trees model

This tutorial teaches you how to use a boosted trees classifier model to predict the income range of individuals based on their demographic data. The model predicts whether a value falls into one of two categories, in this case whether an individual's annual income falls above or below $50,000.

This tutorial uses the bigquery-public-data.ml_datasets.census_adult_income dataset. This dataset contains the demographic and income information of US residents from 2000 and 2010.

Create a dataset

Create a BigQuery dataset to store your ML model.

Console

  1. In the Google Cloud console, go to the BigQuery page.

    Go to the BigQuery page

  2. In the Explorer pane, click your project name.

  3. Click View actions > Create dataset

  4. On the Create dataset page, do the following:

    • For Dataset ID, enter bqml_tutorial.

    • For Location type, select Multi-region, and then select US (multiple regions in United States).

    • Leave the remaining default settings as they are, and click Create dataset.

bq

To create a new dataset, use the bq mk command with the --location flag. For a full list of possible parameters, see the bq mk --dataset command reference.

  1. Create a dataset named bqml_tutorial with the data location set to US and a description of BigQuery ML tutorial dataset:

    bq --location=US mk -d \
 --description "BigQuery ML tutorial dataset." \
 bqml_tutorial

    Instead of using the --dataset flag, the command uses the -d shortcut. If you omit -d and --dataset, the command defaults to creating a dataset.

  2. Confirm that the dataset was created:

    bq ls

API

Call the datasets.insert method with a defined dataset resource.

{
  "datasetReference": {
     "datasetId": "bqml_tutorial"
  }
}

BigQuery DataFrames

Before trying this sample, follow the BigQuery DataFrames setup instructions in the BigQuery quickstart using BigQuery DataFrames. For more information, see the BigQuery DataFrames reference documentation.

To authenticate to BigQuery, set up Application Default Credentials. For more information, see Set up ADC for a local development environment. 

import google.cloud.bigquery

bqclient = google.cloud.bigquery.Client()
bqclient.create_dataset("bqml_tutorial", exists_ok=True)

Prepare the sample data

The model you create in this tutorial predicts the income bracket for census respondents, based on the following features:

  • Age
  • Type of work performed
  • Marital status
  • Level of education
  • Occupation
  • Hours worked per week

The education column isn't included in the training data, because the education and education_num columns both express the respondent's level of education in different formats.

You separate the data into training, evaluation, and prediction sets by creating a new dataframe column that is derived from the functional_weight column. Eighty percent of the data is used for training the model, and the remaining twenty percent of the data is used for evaluation and prediction.

SQL

To prepare your sample data, create a view to contain the training data. This view is used by the CREATE MODEL statement later in this tutorial.

Run the query that prepares the sample data:

  1. In the Google Cloud console, go to the BigQuery page.

    Go to BigQuery

  2. In the query editor, run the following query:

    CREATE OR REPLACE VIEW
  `bqml_tutorial.input_data` AS
SELECT
  age,
  workclass,
  marital_status,
  education_num,
  occupation,
  hours_per_week,
  income_bracket,
  CASE
    WHEN MOD(functional_weight, 10) < 8 THEN 'training'
    WHEN MOD(functional_weight, 10) = 8 THEN 'evaluation'
    WHEN MOD(functional_weight, 10) = 9 THEN 'prediction'
  END AS dataframe
FROM
  `bigquery-public-data.ml_datasets.census_adult_income`;

  3. In the left pane, click Explorer:

    Highlighted button for the Explorer pane.

    If you don't see the left pane, click Expand left pane to open the pane.

  4. In the Explorer pane, search for the bqml_tutorial dataset.

  5. Click the dataset, and then click Overview > Tables.

  6. Click the input_data view to open the information pane. The view schema appears in the Schema tab.

BigQuery DataFrames

Create a DataFrame called input_data. You use input_data later in this tutorial to use to train the model, evaluate it, and make predictions.

Before trying this sample, follow the BigQuery DataFrames setup instructions in the BigQuery quickstart using BigQuery DataFrames. For more information, see the BigQuery DataFrames reference documentation.

To authenticate to BigQuery, set up Application Default Credentials. For more information, see Set up ADC for a local development environment. 

import bigframes.pandas as bpd

input_data = bpd.read_gbq(
    "bigquery-public-data.ml_datasets.census_adult_income",
    columns=(
        "age",
        "workclass",
        "marital_status",
        "education_num",
        "occupation",
        "hours_per_week",
        "income_bracket",
        "functional_weight",
    ),
)
input_data["dataframe"] = bpd.Series("training", index=input_data.index,).case_when(
    [
        (((input_data["functional_weight"] % 10) == 8), "evaluation"),
        (((input_data["functional_weight"] % 10) == 9), "prediction"),
    ]
)
del input_data["functional_weight"]

Create the boosted trees model

Create a boosted trees model to predict census respondents' income bracket, and train it on the census data. The query takes about 30 minutes to complete.

SQL

Follow these steps to create the model:

  1. In the Google Cloud console, go to the BigQuery page.

    Go to BigQuery

  2. In the query editor, paste in the following query and click Run:

    CREATE MODEL `bqml_tutorial.tree_model`
OPTIONS(MODEL_TYPE='BOOSTED_TREE_CLASSIFIER',
        BOOSTER_TYPE = 'GBTREE',
        NUM_PARALLEL_TREE = 1,
        MAX_ITERATIONS = 50,
        TREE_METHOD = 'HIST',
        EARLY_STOP = FALSE,
        SUBSAMPLE = 0.85,
        INPUT_LABEL_COLS = ['income_bracket'])
AS SELECT * EXCEPT(dataframe)
FROM `bqml_tutorial.input_data`
WHERE dataframe = 'training';

    After the query completes, the tree_model model can be accessed through the Explorer pane. Because the query uses a CREATE MODEL statement to create a model, you don't see query results.

BigQuery DataFrames

Before trying this sample, follow the BigQuery DataFrames setup instructions in the BigQuery quickstart using BigQuery DataFrames. For more information, see the BigQuery DataFrames reference documentation.

To authenticate to BigQuery, set up Application Default Credentials. For more information, see Set up ADC for a local development environment. 

from bigframes.ml import ensemble

# input_data is defined in an earlier step.
training_data = input_data[input_data["dataframe"] == "training"]
X = training_data.drop(columns=["income_bracket", "dataframe"])
y = training_data["income_bracket"]

# create and train the model
tree_model = ensemble.XGBClassifier(
    n_estimators=1,
    booster="gbtree",
    tree_method="hist",
    max_iterations=1,  # For a more accurate model, try 50 iterations.
    subsample=0.85,
)
tree_model.fit(X, y)

tree_model.to_gbq(
    your_model_id,  # For example: "your-project.bqml_tutorial.tree_model"
    replace=True,
)

Evaluate the model

SQL

Follow these steps to evaluate the model:

  1. In the Google Cloud console, go to the BigQuery page.

    Go to BigQuery

  2. In the query editor, paste in the following query and click Run:

      SELECT
    *
  FROM
    ML.EVALUATE (MODEL `bqml_tutorial.tree_model`,
      (
      SELECT
        *
      FROM
        `bqml_tutorial.input_data`
      WHERE
        dataframe = 'evaluation'
      )
    );

    The results should look similar to the following:

    +---------------------+---------------------+---------------------+-------------------+---------------------+---------------------+
| precision           | recall              | accuracy            | f1_score          | log_loss            | roc_auc             |
+---------------------+---------------------+---------------------+-------------------+-------------------------------------------+
| 0.67192429022082023 | 0.57880434782608692 | 0.83942963422194672 | 0.621897810218978 | 0.34405456040833338 | 0.88733566433566435 |
+---------------------+---------------------+ --------------------+-------------------+---------------------+---------------------+

BigQuery DataFrames

Before trying this sample, follow the BigQuery DataFrames setup instructions in the BigQuery quickstart using BigQuery DataFrames. For more information, see the BigQuery DataFrames reference documentation.

To authenticate to BigQuery, set up Application Default Credentials. For more information, see Set up ADC for a local development environment. 

# Select model you'll use for predictions. `read_gbq_model` loads model
# data from BigQuery, but you could also use the `tree_model` object
# from the previous step.
tree_model = bpd.read_gbq_model(
    your_model_id,  # For example: "your-project.bqml_tutorial.tree_model"
)

# input_data is defined in an earlier step.
evaluation_data = input_data[input_data["dataframe"] == "evaluation"]
X = evaluation_data.drop(columns=["income_bracket", "dataframe"])
y = evaluation_data["income_bracket"]

# The score() method evaluates how the model performs compared to the
# actual data. Output DataFrame matches that of ML.EVALUATE().
score = tree_model.score(X, y)
score.peek()
# Output:
#    precision    recall  accuracy  f1_score  log_loss   roc_auc
# 0   0.671924  0.578804  0.839429  0.621897  0.344054  0.887335

The evaluation metrics indicate good model performance, in particular, the fact that the roc_auc score is greater than 0.8.

For more information about the evaluation metrics, see Output.

Use the model to predict classifications

SQL

Follow these steps to forecast data with the model:

  1. In the Google Cloud console, go to the BigQuery page.

    Go to BigQuery

  2. In the query editor, paste in the following query and click Run:

      SELECT
    *
  FROM
    ML.PREDICT (MODEL `bqml_tutorial.tree_model`,
      (
      SELECT
        *
      FROM
        `bqml_tutorial.input_data`
      WHERE
        dataframe = 'prediction'
      )
    );

The first few columns of the results should look similar to the following:

  +---------------------------+--------------------------------------+-------------------------------------+
  | predicted_income_bracket  | predicted_income_bracket_probs.label | predicted_income_bracket_probs.prob |
  +---------------------------+--------------------------------------+-------------------------------------+
  |  <=50K                    |  >50K                                | 0.05183430016040802                 |
  +---------------------------+--------------------------------------+-------------------------------------+
  |                           |  <50K                                | 0.94816571474075317                 |
  +---------------------------+--------------------------------------+-------------------------------------+
  |  <=50K                    |  >50K                                | 0.00365859130397439                 |
  +---------------------------+--------------------------------------+-------------------------------------+
  |                           |  <50K                                | 0.99634140729904175                 |
  +---------------------------+--------------------------------------+-------------------------------------+
  |  <=50K                    |  >50K                                | 0.037775970995426178                |
  +---------------------------+--------------------------------------+-------------------------------------+
  |                           |  <50K                                | 0.96222406625747681                 |
  +---------------------------+--------------------------------------+-------------------------------------+

BigQuery DataFrames

Before trying this sample, follow the BigQuery DataFrames setup instructions in the BigQuery quickstart using BigQuery DataFrames. For more information, see the BigQuery DataFrames reference documentation.

To authenticate to BigQuery, set up Application Default Credentials. For more information, see Set up ADC for a local development environment. 

# Select model you'll use for predictions. `read_gbq_model` loads model
# data from BigQuery, but you could also use the `tree_model` object
# from previous steps.
tree_model = bpd.read_gbq_model(
    your_model_id,  # For example: "your-project.bqml_tutorial.tree_model"
)

# input_data is defined in an earlier step.
prediction_data = input_data[input_data["dataframe"] == "prediction"]

predictions = tree_model.predict(prediction_data)
predictions.peek()
# Output:
# predicted_income_bracket   predicted_income_bracket_probs.label  predicted_income_bracket_probs.prob
#                   <=50K                                   >50K                   0.05183430016040802
#                                                           <50K                   0.94816571474075317
#                   <=50K                                   >50K                   0.00365859130397439
#                                                           <50K                   0.99634140729904175
#                   <=50K                                   >50K                   0.037775970995426178
#                                                           <50K                   0.96222406625747681

The predicted_income_bracket contains the predicted value from the model. The predicted_income_bracket_probs.label shows the two labels that the model had to choose between, and the predicted_income_bracket_probs.prob column shows the probability of the given label being the correct one.

For more information about the output columns, see Classification models.