Google Search Console API Integration

This handler integrates with the Google Search API to allow you to use Google Search data in your SQL queries.

Example: Predicting clicks based on Google Search activity

To see how the Google Search Console handler is used, let's walk through a simple example to create a model to predict our clicks on a site based on Google Search activity.

Connect to the Google Search API

We start by creating a database to connect to the Google Search Console API. Currently, there is no need for an API key:

However, you will need to have a Google account and have enabled the Google Search Console API. Also, you will need to have the credentials in a json file. You can find more information on how to do this here.

Optional: The credentials file can be stored in the google_search handler folder in the mindsdb/integrations/google_search_handler directory.

CREATE
DATABASE my_Search
WITH  ENGINE = 'google_search',
parameters = {
    'credentials': '/path/to/credentials.json'
};    

This creates a database called my_Search. This database ships with a table called Analytics and with a table called Sitemaps that we can use to search for Google Search data as well as to process Google Search data.

Query your search traffic data with filters and parameters that you define

Let's get traffic data for a specific site.

SELECT clicks
FROM my_console.Analytics
WHERE siteUrl = 'https://www.mindsdb.com'
  AND startDate = '2020-10-01'
  AND endDate = '2020-10-31'
  AND dimensions = 'query'
  AND type = 'web'
GROUP BY query
ORDER BY clicks

Submit a sitemap to Google Search Console

Let's test by submitting a sitemap to Google Search Console.

INSERT INTO my_console.Sitemaps (siteUrl, sitemapUrl)
VALUES ('https://www.mindsdb.com', 'https://www.mindsdb.com/sitemap.xml')

Delete a sitemap from Google Search Console

Let's test by deleting a sitemap from Google Search Console.

DELETE FROM my_console.Sitemaps
WHERE siteUrl = 'https://www.mindsdb.com'
  AND feedpath = 'https://www.mindsdb.com/sitemap.xml'

Creating a model to predict future clicks

Now we can incorporate external data from Google Search within our ML models.

CREATE
PREDICTOR my_search_clicks
FROM my_search.Analytics
PREDICT
clicks