Create machine learning models in BigQuery ML#

https://cloud.google.com/bigquery-ml/docs/create-machine-learning-model

This tutorial introduces users to BigQuery ML using the Google Cloud console.

BigQuery ML enables users to create and execute machine learning models in BigQuery by using SQL queries. The goal is to democratize machine learning by enabling SQL practitioners to build models using their existing tools and to increase development speed by eliminating the need for data movement.

In this tutorial, you use the sample Google Analytics sample dataset for BigQuery to create a model that predicts whether a website visitor will make a transaction. For information on the schema of the Analytics dataset, see BigQuery export schema in the Google Analytics Help Center.

Objectives#

In this tutorial, you use:

BigQuery ML to create a binary logistic regression model using the CREATE MODEL statement
The ML.EVALUATE function to evaluate the ML model
The ML.PREDICT function to make predictions using the ML model

Costs#

This tutorial uses billable components of Google Cloud, including the following:

BigQuery
BigQuery ML For more information on BigQuery costs, see the BigQuery pricing page.

For more information on BigQuery ML costs, see the BigQuery ML pricing page.

import os

print(os.environ.get("GOOGLE_CLOUD_PROJECT"))

ambient-skyline-390606

from google.cloud import bigquery

bigquery_client = bigquery.Client()
%load_ext google.cloud.bigquery

Create your dadaset#

The first step is to create a BigQuery dataset to store your ML model. To create your dataset:

dataset = bigquery_client.create_dataset("bqml_tutorial", exists_ok=True)

Create your model#

Next, you create a logistic regression model using the Google Analytics sample dataset for BigQuery. The following standard SQL query is used to create the model you use to predict whether a website visitor will make a transaction.

CREATE OR REPLACE MODEL `bqml_tutorial.sample_model`
OPTIONS(model_type='logistic_reg') AS
SELECT
  IF(totals.transactions IS NULL, 0, 1) AS label,
  IFNULL(device.operatingSystem, "") AS os,
  device.isMobile AS is_mobile,
  IFNULL(geoNetwork.country, "") AS country,
  IFNULL(totals.pageviews, 0) AS pageviews
FROM
  `bigquery-public-data.google_analytics_sample.ga_sessions_*`
WHERE
  _TABLE_SUFFIX BETWEEN '20160801' AND '20170630'

In addition to creating the model, running a query that contains the CREATE MODEL statement trains the model using the data retrieved by your query’s SELECT statement.

Query details#

The CREATE MODEL clause is used to create and train the model named bqml_tutorial.sample_model.

The OPTIONS(model_type='logistic_reg') clause indicates that you are creating a logistic regression model. A logistic regression model tries to split input data into two classes and gives the probability the data is in one of the classes. Usually, what you are trying to detect (such as whether an email is spam) is represented by 1 and everything else is represented by 0. If the logistic regression model outputs 0.9, there is a 90% probability the input is what you are trying to detect (the email is spam).

This query’s SELECT statement retrieves the following columns that are used by the model to predict the probability a customer will complete a transaction:

totals.transactions — The total number of ecommerce transactions within the session. If the number of transactions is NULL, the value in the label column is set to 0. Otherwise, it is set to 1. These values represent the possible outcomes. Creating an alias named label is an alternative to setting the input_label_cols= option in the CREATE MODEL statement.
device.operatingSystem — The operating system of the visitor’s device.
device.isMobile — Indicates whether the visitor’s device is a mobile device.
geoNetwork.country — The country from which the sessions originated, based on the IP address.
totals.pageviews — The total number of page views within the session.

The FROM clause — bigquery-public-data.google_analytics_sample.ga_sessions_* — indicates that you are querying the Google Analytics sample dataset. This dataset is in the bigquery-public-data project. You are querying a set of tables sharded by date. This is represented by the wildcard in the table name: google_analytics_sample.ga_sessions_*.

The WHERE clause — _TABLE_SUFFIX BETWEEN '20160801' AND '20170630' — limits the number of tables scanned by the query. The date range scanned is August 1, 2016 to June 30, 2017.

Run the CREATE MODEL query#

%%bigquery
CREATE OR REPLACE MODEL `bqml_tutorial.sample_model`
OPTIONS(model_type='logistic_reg') AS
SELECT
  IF(totals.transactions IS NULL, 0, 1) AS label,
  IFNULL(device.operatingSystem, "") AS os,
  device.isMobile AS is_mobile,
  IFNULL(geoNetwork.country, "") AS country,
  IFNULL(totals.pageviews, 0) AS pageviews
FROM
  `bigquery-public-data.google_analytics_sample.ga_sessions_*`
WHERE
  _TABLE_SUFFIX BETWEEN '20160801' AND '20170630'

The query takes several minutes to complete. After the first iteration is complete, your model (sample_model) appears in the navigation panel. Because the query uses a CREATE MODEL statement to create a model, you do not see query results.

You can observe the model as it’s being trained by viewing the Model stats tab. As soon as the first iteration completes, the tab is updated. The stats continue to update as each iteration completes.

Get training statistics#

To see the results of the model training, you can use the ML.TRAINING_INFO function, or you can view the statistics in the Google Cloud console.

Machine learning is about creating a model that can use data to make a prediction. The model is essentially a function that takes inputs and applies calculations to the inputs to produce an output — a prediction.

Machine learning algorithms work by taking several examples where the prediction is already known (such as the historical data of user purchases) and iteratively adjusting various weights in the model so that the model’s predictions match the true values. It does this by minimizing how wrong the model is using a metric called loss.

The expectation is that for each iteration, the loss should be decreasing (ideally to zero). A loss of zero means the model is 100% accurate.

To see the model training statistics that were generated when you ran the CREATE MODEL query:

%%bigquery
SELECT
  *
FROM
  ML.TRAINING_INFO(MODEL `bqml_tutorial.sample_model`)

	iteration	loss	eval_loss	learning_rate	duration_ms
0	8	0.043878	0.045445	25.6	10029
1	7	0.044654	0.045499	25.6	10281
2	6	0.047345	0.048273	12.8	10063
3	5	0.053888	0.053337	6.4	11109
4	4	0.067776	0.066406	3.2	11236
5	3	0.097545	0.096203	1.6	9869
6	2	0.169802	0.168849	0.8	10523
7	1	0.320692	0.320174	0.4	12159
8	0	0.521573	0.521380	0.2	8892

The Training Data Loss column represents the loss metric calculated after the given iteration on the training dataset. Since you performed a logistic regression, this column is the log loss. The Evaluation Data Loss column is the same loss metric calculated on the holdout dataset (data that is held back from training to validate the model).

BigQuery ML automatically splits your input data into a training set and a holdout set to avoid overfitting the model. This is necessary so that the training algorithm doesn’t so closely tailor to the known data that it doesn’t generalize to unseen, new examples.

Training Data Loss and Evaluation Data Loss are average loss values, averaged over all examples in the respective sets.

For more details on the ML.TRAINING_INFO function, see the BigQuery ML Syntax Reference.

Evaluate your model#

After creating your model, you evaluate the performance of the classifier using the ML.EVALUATE function. The ML.EVALUATE function evaluates the predicted values against the actual data. You can also use the ML.ROC_CURVE function for logistic regression specific metrics.

In this tutorial you are using a binary classification model that detects transactions. The two classes are the values in the label column: 0 (no transactions) and 1 (transaction made).

The query used to evaluate the model is as follows:

SELECT
  *
FROM
  ML.EVALUATE(MODEL `bqml_tutorial.sample_model`, (
SELECT
  IF(totals.transactions IS NULL, 0, 1) AS label,
  IFNULL(device.operatingSystem, "") AS os,
  device.isMobile AS is_mobile,
  IFNULL(geoNetwork.country, "") AS country,
  IFNULL(totals.pageviews, 0) AS pageviews
FROM
  `bigquery-public-data.google_analytics_sample.ga_sessions_*`
WHERE
  _TABLE_SUFFIX BETWEEN '20170701' AND '20170801'))

Query details#

The top-most SELECT statement retrieves the columns from your model.

The FROM clause uses the ML.EVALUATE function against your model: bqml_tutorial.sample_model.

This query’s nested SELECT statement and FROM clause are the same as those in the CREATE MODEL query.

The WHERE clause — _TABLE_SUFFIX BETWEEN '20170701' AND '20170801' — limits the number of tables scanned by the query. The date range scanned is July 1, 2017 to August 1, 2017. This is the data you’re using to evaluate the predictive performance of the model. It was collected in the month immediately following the time period spanned by the training data.

Run the ML.EVALUATE query#

%%bigquery
SELECT
  *
FROM
  ML.EVALUATE(MODEL `bqml_tutorial.sample_model`, (
SELECT
  IF(totals.transactions IS NULL, 0, 1) AS label,
  IFNULL(device.operatingSystem, "") AS os,
  device.isMobile AS is_mobile,
  IFNULL(geoNetwork.country, "") AS country,
  IFNULL(totals.pageviews, 0) AS pageviews
FROM
  `bigquery-public-data.google_analytics_sample.ga_sessions_*`
WHERE
  _TABLE_SUFFIX BETWEEN '20170701' AND '20170801'))

	precision	recall	accuracy	f1_score	log_loss	roc_auc
0	0.468504	0.110801	0.985343	0.179217	0.046242	0.981741

Because you performed a logistic regression, the results include the following columns:

precision — A metric for classification models. Precision identifies the frequency with which a model was correct when predicting the positive class.
recall — A metric for classification models that answers the following question: Out of all the possible positive labels, how many did the model correctly identify?
accuracy — Accuracy is the fraction of predictions that a classification model got right.
f1_score — A measure of the accuracy of the model. The f1 score is the harmonic average of the precision and recall. An f1 score’s best value is 1. The worst value is 0.
log_loss — The loss function used in a logistic regression. This is the measure of how far the model’s predictions are from the correct labels.
roc_auc — The area under the ROC curve. This is the probability that a classifier is more confident that a randomly chosen positive example is actually positive than that a randomly chosen negative example is positive. For more information, see Classification in the Machine Learning Crash Course.

Use your model to predict outcomes#

Now that you have evaluated your model, the next step is to use it to predict an outcome. You use your model to predict the number of transactions made by website visitors from each country.

The query used to predict the outcome is as follows:

SELECT
  country,
  SUM(predicted_label) as total_predicted_purchases
FROM
  ML.PREDICT(MODEL `bqml_tutorial.sample_model`, (
SELECT
  IFNULL(device.operatingSystem, "") AS os,
  device.isMobile AS is_mobile,
  IFNULL(totals.pageviews, 0) AS pageviews,
  IFNULL(geoNetwork.country, "") AS country
FROM
  `bigquery-public-data.google_analytics_sample.ga_sessions_*`
WHERE
  _TABLE_SUFFIX BETWEEN '20170701' AND '20170801'))
GROUP BY country
ORDER BY total_predicted_purchases DESC
LIMIT 10

Query details#

The top-most SELECT statement retrieves the country column and sums the predicted_label column. This column is generated by the ML.PREDICT function. When you use the ML.PREDICT function the output column name for the model is predicted_<label_column_name>. For linear regression models, predicted_label is the estimated value of label. For logistic regression models, predicted_label is the most likely label, which in this case is either 0 or 1.

The ML.PREDICT function is used to predict results using your model: bqml_tutorial.sample_model.

This query’s nested SELECT statement and FROM clause are the same as those in the CREATE MODEL query.

The WHERE clause — _TABLE_SUFFIX BETWEEN '20170701' AND '20170801' — limits the number of tables scanned by the query. The date range scanned is July 1, 2017 to August 1, 2017. This is the data for which you’re making predictions. It was collected in the month immediately following the time period spanned by the training data.

The GROUP BY and ORDER BY clauses group the results by country and order them by the sum of the predicted purchases in descending order.

The LIMIT clause is used here to display only the top 10 results.

Run the ML.PREDICT query#

%%bigquery
SELECT
  country,
  SUM(predicted_label) as total_predicted_purchases
FROM
  ML.PREDICT(MODEL `bqml_tutorial.sample_model`, (
SELECT
  IFNULL(device.operatingSystem, "") AS os,
  device.isMobile AS is_mobile,
  IFNULL(totals.pageviews, 0) AS pageviews,
  IFNULL(geoNetwork.country, "") AS country
FROM
  `bigquery-public-data.google_analytics_sample.ga_sessions_*`
WHERE
  _TABLE_SUFFIX BETWEEN '20170701' AND '20170801'))
GROUP BY country
ORDER BY total_predicted_purchases DESC
LIMIT 10

	country	total_predicted_purchases
0	United States	220
1	Taiwan	8
2	Canada	7
3	Turkey	2
4	Japan	2
5	India	2
6	Indonesia	1
7	Venezuela	1
8	Italy	1
9	St. Lucia	1

Predict purchases per user#

In this example, you try to predict the number of transactions each website visitor will make. This query is identical to the previous query except for the GROUP BY clause. Here the GROUP BY clause — GROUP BY fullVisitorId — is used to group the results by visitor ID.

%%bigquery
SELECT
  fullVisitorId,
  SUM(predicted_label) as total_predicted_purchases
FROM
  ML.PREDICT(MODEL `bqml_tutorial.sample_model`, (
SELECT
  IFNULL(device.operatingSystem, "") AS os,
  device.isMobile AS is_mobile,
  IFNULL(totals.pageviews, 0) AS pageviews,
  IFNULL(geoNetwork.country, "") AS country,
  fullVisitorId
FROM
  `bigquery-public-data.google_analytics_sample.ga_sessions_*`
WHERE
  _TABLE_SUFFIX BETWEEN '20170701' AND '20170801'))
GROUP BY fullVisitorId
ORDER BY total_predicted_purchases DESC
LIMIT 10

	fullVisitorId	total_predicted_purchases
0	9417857471295131045	4
1	057693500927581077	2
2	2158257269735455737	2
3	5073919761051630191	2
4	8388931032955052746	2
5	7420300501523012460	2
6	112288330928895942	2
7	0456807427403774085	2
8	2105122376016897629	2
9	489038402765684003	2

Clean up#

To avoid incurring charges to your Google Cloud account for the resources used on this page, follow these steps.

You can delete the project you created.
Or you can keep the project and delete the dataset.

Delete your dasaet#

Deleting your project removes all datasets and all tables in the project. If you prefer to reuse the project, you can delete the dataset you created in this tutorial:

bigquery_client.delete_dataset(dataset, delete_contents=True)

Eduardo Avelar

Create machine learning models in BigQuery ML

Contents

Create machine learning models in BigQuery ML#

Objectives#

Costs#

Create your dadaset#

Create your model#

Query details#

Run the CREATE MODEL query#

Get training statistics#

Evaluate your model#

Query details#

Run the ML.EVALUATE query#

Use your model to predict outcomes#

Query details#

Run the ML.PREDICT query#

Predict purchases per user#

Clean up#

Delete your dasaet#