Use BigQuery ML to make recommendations from Google analytics data

Step one: Create your dataset

Create a BigQuery dataset to store your ML model:

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 more_vert 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).

     The public datasets are stored in the US multi-region. For simplicity, store your dataset in the same location.

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

     

Step two: Load Analytics 360 data into BigQuery

Most of the time, the ratings in your data will not reflect a value that a user sets explicitly. In these scenarios, we can devise a proxy to these values as an implicit rating and use a different algorithm to compute recommendations. In this sample, we will a sample Analytics 360 dataset. This sample is based on the following article.

The following is a query to run to create a dataset with implicit ratings from the session duration that a visitor has had on a page from the cloud-training-demos.GA360_test.ga_sessions_sample. The goal of this query is to create a dataset with three columns that we can map to a user column, item column and rating column.

1. In the Google Cloud console, click the Compose new query button.

2. Enter the following GoogleSQL query in the Query editor text area.

   #standardSQL
   CREATE OR REPLACE TABLE
    bqml_tutorial.analytics_session_data AS
   WITH
    visitor_page_content AS (
    SELECT
      fullVisitorID,
      (
      SELECT
        MAX(
        IF
          (index=10,
            value,
            NULL))
      FROM
        UNNEST(hits.customDimensions)) AS latestContentId,
      (LEAD(hits.time, 1)
        OVER (PARTITION BY fullVisitorId ORDER BY hits.time ASC) - hits.time)
                AS session_duration
    FROM
      `cloud-training-demos.GA360_test.ga_sessions_sample`,
      UNNEST(hits) AS hits
    WHERE
      # only include hits on pages
      hits.type = "PAGE"
    GROUP BY
      fullVisitorId,
      latestContentId,
      hits.time )
    # aggregate web stats
   SELECT
    fullVisitorID AS visitorId,
    latestContentId AS contentId,
    SUM(session_duration) AS session_duration
   FROM
    visitor_page_content
   WHERE
    latestContentId IS NOT NULL
   GROUP BY
    fullVisitorID,
    latestContentId
   HAVING
    session_duration > 0
   ORDER BY
    latestContentId

3. (Optional) To set the processing location, click More > Query settings. For Processing location, choose US. This step is optional because the processing location is automatically detected based on the dataset's location.

   

4. Click Run.

   When the query finishes running, (bqml_tutorial.analytics_session_data) will appear in the navigation panel. Because the query uses a CREATE TABLE statement to create a table, you do not see the query results.

5. If you take a look at the table produced it should look something like the following:

   

   Note that this result is specific to how the data was exported to BigQuery. The query to extract your own data might differ.

Step three: Create your Implicit Recommendations Model

Next, you create an implicit recommendations model using the google analytics table that was loaded in the previous step. The following GoogleSQL query is used to create the model that will be used to predict a confidence rating for every visitorId contentId pair. A rating is created with centering and scaling by the median session duration, and filtering those records where the session duration is more than 3.33 times the median as outliers.

#standardSQL
CREATE OR REPLACE MODEL `bqml_tutorial.my_implicit_mf_model`
OPTIONS
  (model_type='matrix_factorization',
   feedback_type='implicit',
   user_col='visitorId',
   item_col='contentId',
   rating_col='rating',
   l2_reg=30,
   num_factors=15) AS
SELECT
  visitorId,
  contentId,
  0.3 * (1 + (session_duration - 57937) / 57937) AS rating
FROM `bqml_tutorial.analytics_session_data`
WHERE 0.3 * (1 + (session_duration - 57937) / 57937) < 1

Query details

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

The OPTIONS(model_type='matrix_factorization', feedback_type='IMPLICIT', user_col='visitorId', ...) clause indicates that you are creating a matrix factorization model. Because feedback_type='IMPLICIT' is specified an implicit matrix factorization model will be trained. An example of how to create an explicit matrix factorization model is explained in Creating an explicit matrix factorization model.

This query's SELECT statement uses the following columns to generate recommendations.

• visitorId—The visitor ID (INT64).
• contentId—The content ID (INT64).
• rating—The implicit rating from 0 to 1 calculated for visitorId and their respective contentId centered and scaled (FLOAT64).

The FROM clause—bqml_tutorial.analytics_session_data — indicates that you are querying the analytics_session_data table in the bqml_tutorial dataset. This dataset is in your bigquery project if the instructions in step two and eight were followed.

Run the CREATE MODEL query

To run the CREATE MODEL query to create and train your model:

1. In the Google Cloud console, click the Compose new query button.

2. Enter the following GoogleSQL query in the Query editor text area.

   #standardSQL
   CREATE OR REPLACE MODEL `bqml_tutorial.my_implicit_mf_model`
   OPTIONS
    (model_type='matrix_factorization',
     feedback_type='implicit',
     user_col='visitorId',
     item_col='contentId',
     rating_col='rating',
     l2_reg=30,
     num_factors=15) AS
   SELECT
    visitorId,
    contentId,
    0.3 * (1 + (session_duration - 57937) / 57937) AS rating
   FROM `bqml_tutorial.analytics_session_data`

3. Click Run.

   The query takes about 12 minutes to complete, after which your model (my_implicit_mf_model) appears in the navigation panel. Because the query uses a CREATE MODEL statement to create a model, you do not see query results.

Step four (optional): 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. In this tutorial, you use the Google Cloud console.

A machine learning algorithm builds a model by examining many examples and attempting to find a model that minimizes loss. This process is called empirical risk minimization.

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

1. In the Google Cloud console navigation panel, in the Resources section, expand [PROJECT_ID] > bqml_tutorial and then click my_implicit_mf_model.

2. Click the Training tab, and then click Table. The results should look like the following:

   

   The Training Data Loss column represents the loss metric calculated after the model is trained on the training dataset. Since you performed matrix factorization, this column is the mean squared error. By default, matrix factorization models will not split the data, so the Evaluation Data Loss column will not be present unless a holdout dataset is specified because splitting the data has a chance of losing all the ratings for a user or an item. As a result, the model will not have latent factor information about missing users or items.

   For more details on the ML.TRAINING_INFO function, see the BigQuery ML syntax reference.

Step five: Evaluate your model

After creating your model, you evaluate the performance of the recommender using the ML.EVALUATE function. The ML.EVALUATE function evaluates the predicted ratings against the actual ratings.

The query used to evaluate the model is as follows:

#standardSQL
SELECT
  *
FROM
  ML.EVALUATE(MODEL `bqml_tutorial.my_implicit_mf_model`)

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.my_implicit_mf_model.

Run the ML.EVALUATE query

To run the ML.EVALUATE query that evaluates the model:

1. In the Google Cloud console, click the Compose new query button.

2. Enter the following GoogleSQL query in the Query editor text area.

   #standardSQL
   SELECT
    *
   FROM
    ML.EVALUATE(MODEL `bqml_tutorial.my_implicit_mf_model`)

3. (Optional) To set the processing location, click More > Query settings. For Processing location, choose US. This step is optional because the processing location is automatically detected based on the dataset's location.

   

4. Click Run.

5. When the query is complete, click the Results tab below the query text area. The results should look like the following:

   

   Because you performed an implicit matrix factorization, the results include the following columns:

   • mean_average_precision
   • mean_squared_error
   • normalized_discounted_cumulative_gain
   • average_rank

   mean_average_precision, normalized_discounted_cumulative_gain, and average_rank are ranking metrics that are explained here: Implicit matrix factorization metrics

Step six: Predict ratings and make recommendations

Use your model to predict ratings and make recommendations

Find all the contentId rating confidences for a set of visitorIds

ML.RECOMMEND does not need to take any additional arguments besides the model, but can take in an optional table. If the input table only has one column that matches the name of the input user or input item column, then all of the predicted item ratings for each user will be outputted and vice versa. Note that if all of the users or all of the items are in the input table, it will output the same results as passing no optional argument to ML.RECOMMEND.

The following is an example of a query to fetch all of the predicted rating confidences for 5 visitors.

#standardSQL
SELECT
  *
FROM
  ML.RECOMMEND(MODEL `bqml_tutorial.my_implicit_mf_model`,
    (
    SELECT
      visitorId
    FROM
      `bqml_tutorial.analytics_session_data`
    LIMIT 5))

Query details

The top-most SELECT statement retrieves the visitorId, contentId, and predicted_rating_confidence column. This last column is generated by the ML.RECOMMEND function. When you use the ML.RECOMMEND function, the output column name for implicit matrix factorization models is predicted_rating-column-name_confidence. For implicit matrix factorization models, predicted_rating_confidence is the estimated confidence for the user/item pair. This confidence value approximately lies between 0 and 1 where the higher confidence indicates that the user prefers item more than an item with a lower confidence value.

The ML.RECOMMEND function is used to predict ratings using your model: bqml_tutorial.my_implicit_mf_model.

This query's nested SELECT statement selects just the visitorId column from the original table used for training.

The LIMIT clause—LIMIT 5—will randomly filter out 5 visitorIds to send to ML.RECOMMEND.

Find the ratings for all visitorId contentId pairs

Now that you have evaluated your model, the next step is to use it to get a rating confidence. You use your model to predict the confidences of every user-item combination in the following query:

#standardSQL
SELECT
  *
FROM
  ML.RECOMMEND(MODEL `bqml_tutorial.my_implicit_mf_model`)

Query details

The top-most SELECT statement retrieves the visitorId, contentId, and predicted_rating_confidence column. This last column is generated by the ML.RECOMMEND function. When you use the ML.RECOMMEND function, the output column name for implicit matrix factorization models is predicted_rating-column-name_confidence. For implicit matrix factorization models, predicted_rating_confidence is the estimated confidence for the user/item pair. This confidence value approximately lies between 0 and 1 where the higher confidence indicates that the user prefers item more than an item with a lower confidence value.

The ML.RECOMMEND function is used to predict ratings using your model: bqml_tutorial.my_implicit_mf_model.

One way of saving the result to table is:

#standardSQL
CREATE OR REPLACE TABLE `bqml_tutorial.recommend_content`
OPTIONS() AS
SELECT
  *
FROM
  ML.RECOMMEND(MODEL `bqml_tutorial.my_implicit_mf_model`)

If a Query Exceeded Resource Limits error occurs for ML.RECOMMEND, retry with a higher billing tier. In the BigQuery command-line tool, this can be set using --maximum_billing_tier.

Generate recommendations

The following query uses ML.RECOMMEND to output the top 5 recommended contentIds per visitorId.

#standardSQL
SELECT
  visitorId,
  ARRAY_AGG(STRUCT(contentId, predicted_rating_confidence)
    ORDER BY predicted_rating_confidence DESC LIMIT 5) AS rec
FROM
  `bqml_tutorial.recommend_content`
GROUP BY
  visitorId

Query details

The SELECT statement aggregates the results from the ML.RECOMMEND query by using GROUP BY visitorId to aggregate the contentId and predicted_rating_confidence in descending order and only keeps the top 5 content ids.

Using the previous recommendations query, we can order by the predicted rating and output the top predicted items for each user. The following query joins the item_ids with the movie_ids found in the movielens.movie_titles table uploaded earlier and outputs the top 5 recommended movies per user.

Run the ML.RECOMMEND query

To run the ML.RECOMMEND query that outputs the top 5 recommended content ids per visitor id:

1. In the Google Cloud console, click the Compose new query button.

2. Enter the following GoogleSQL query in the Query editor text area.

   #standardSQL
   CREATE OR REPLACE TABLE `bqml_tutorial.recommend_content`
   OPTIONS() AS
   SELECT
    *
   FROM
    ML.RECOMMEND(MODEL `bqml_tutorial.my_implicit_mf_model`)

3. Click Run.

   When the query finishes running, (bqml_tutorial.recommend_content) will appear in the navigation panel of the Google Cloud console. Because the query uses a CREATE TABLE statement to create a table, you do not see the query results.

4. Compose another new query. Enter the following GoogleSQL query in the Query editor text area once the previous query has finished running.

   #standardSQL
   SELECT
    visitorId,
    ARRAY_AGG(STRUCT(contentId, predicted_rating_confidence)
      ORDER BY predicted_rating_confidence DESC LIMIT 5) AS rec
   FROM
    `bqml_tutorial.recommend_content`
   GROUP BY
    visitorId

5. (Optional) To set the processing location, click More > Query settings. For Processing location, choose US. This step is optional because the processing location is automatically detected based on the dataset's location.

   

6. Click Run.

7. When the query is complete, click the Results tab below the query text area. The results should look like the following: