Google Analytics Universal

Google Analytics was a web analytics service offered by Google that tracked and reported website traffic (now replaced with Google Analytics 4). This telescope is deprecated and no longer supported by BAD, however its historical data is still used by the ONIX Workflow.

This telescope obtains data from Google Analytics for one view id per publisher and for several combinations of metrics and dimensions. It is possible to add a regex expression to filter on pagepaths, so only data on relevant pagepaths is collected. Note that Google Analytics data is only available for the last 26 months, see Data retention - Analytics Help for more info.

To get access to the Google Analytics data a publisher, as a user, needs to add the relevant google service account.

Dataset Name	google
Table Names	google_analytics
Table Type	Partitioned
Average Runtime	10 min
Average Download Size	10-20 MB
Harvest Type	API
Run Schedule	Monthly
Catch-up Missed Runs	✅
Each Run Includes All Data	❌

Custom dimensions for ANU Press

ANU Press was using custom dimensions in their Google Analytics data. To ensure that the telescope processes these custom dimensions, the organisation name needs to be set to exactly 'ANU Press'. The organisation name is used directly inside the telescope and if it matches 'ANU Press' additional dimensions are added and a different BigQuery schema is used.

A note on the API metrics

We use the python client for the The Google Analytics API in order to retrieve the data on several metrics (such as page views) per country. It appears as though the API does not return a result for every country. We would have expected any data without a country field to be labelled with a country name of "not set", however this does not appear to be the case. At this time, we have no other way of retrieving country-level data on the desired metrics, so we must acknowledge that the numbers returned by the API are slightly different to those found on the Google Analytics web page. A support ticket has been created with google in the hope of resolving this issue.

Telescope kwargs

Organisation Name (organisation_name)

The name of the organisation as displayed on Google Analytics

View ID (view_id)

The View ID points to the specific view on which Google Analytics data is collected. See the google support page for more information on the hierarchy of the Analytics account.

Pagepath Regex (pagepath_regex)

This is a regular expression that is used to filter on pagepaths for which analytics data is collected. The regular expression can be set to an empty string if no filtering is required. Note that the Google Analytics API uses 're2', so it is not possible to use e.g. negative lookaheads. See the google support page and github wiki for more information.

Setting up service account

• Create a service account from IAM & Admin - Service Accounts

• Create a JSON key and download the file with key

• For each organisation/publisher of interest, ask them to add this service account as a user for the correct view id

Getting the view ID (after given access)

from googleapiclient.discovery import build
from oauth2client.service_account import ServiceAccountCredentials

scopes = ['https://www.googleapis.com/auth/analytics.readonly']
credentials_path = '/path/to/service_account_credentials.json'

creds = ServiceAccountCredentials.from_json_keyfile_name(credentials_path, scopes=scopes)

# Build the service object.
service = build('analytics', 'v3', credentials=creds)

account_summaries = service.management().accountSummaries().list().execute()
view_ids = []
for account in account_summaries['items']:
    account_name = account['name']
    profiles = account['webProperties'][0]['profiles']
    website_url = account['webProperties'][0]['websiteUrl']
    for profile in profiles:
        view_id_info = {'account': account_name, 'websiteUrl': website_url, 'view_id': profile['id'], 
                        'view_name': profile['name']}
        view_ids.append(view_id_info)

Airflow connections

Name	Description
oaebu_service_account	The credentials for the service account that has been given access to the google analytics view.

Telescope Tasks

Data Download & Transform

This downloads a single month of reporting data using the GA3 analytics reporting API. It transforms the data from the report structure into the schema structure and saves it to a .jsonl file.

Big Query Load

The transformed data is loaded from the Google Cloud bucket into a partitioned BigQuery table under the google dataset (which will be created should it not exist yet). Since the data is partitioned on the release month, there will only be a single table named google_analytics3.

Table Schema

Name	Type	Mode	Description
url	STRING	REQUIRED	Base URL of the book pages.
title	STRING	REQUIRED	Title o f the book.
start_date	DATE	REQUIRED	Start date for period of analytics info.
end_date	DATE	REQUIRED	End date for period of analytics info.
average_time	FLOAT	REQUIRED	Average time (in seconds) spent on each page.
unique_views	RECORD	NULLABLE	Unique views for several different dimensions. Unique views is the number of sessions during which the specified page was viewed at least once. A unique pageview is counted for each page URL + page title combination.
unique_views.country	RECORD	REPEATED	Unique views per users' country, derived from their IP addresses or Geographical IDs.
unique_views.country.name	STRING	NULLABLE	Country name.
unique_views.country.value	INTEGER	NULLABLE	Number of unique views.
unique_views.referrer	RECORD	REPEATED	Unique views per referrer, the full referring URL including the hostname and path.
unique_views.referrer.name	STRING	NULLABLE	Referrer name.
unique_views.referrer.value	INTEGER	NULLABLE	Number of unique views.
unique_views.social_network	RECORD	REPEATED	Unique views per social network. This is related to the referring social network for traffic sources; e.g., Google+, Blogger.
unique_views.social_network.name	STRING	NULLABLE	Social network name.
unique_views.social_network.value	INTEGER	NULLABLE	Number of unique views.
page_views	RECORD	NULLABLE	The total number of pageviews for the property
page_views.country	RECORD	REPEATED	Page views per users' country, derived from their IP addresses or Geographical IDs.
page_views.country.name	STRING	NULLABLE	Country name.
page_views.country.value	INTEGER	NULLABLE	Number of page views.
page_views.referrer	RECORD	REPEATED	Page views per referrer, the full referring URL including the hostname and path.
page_views.referrer.name	STRING	NULLABLE	Referrer name.
page_views.referrer.value	INTEGER	NULLABLE	Number of page views.
page_views.social_network	RECORD	REPEATED	Page views per social network. This is related to the referring social network for traffic sources; e.g., Google+, Blogger.
page_views.social_network.name	STRING	NULLABLE	Social network name.
page_views.social_network.value	INTEGER	NULLABLE	Number of page views.
sessions	RECORD	NULLABLE	Total number of sessions for several different dimensions.
sessions.country	RECORD	REPEATED	Unique views per users' country, derived from their IP addresses or Geographical IDs.
sessions.country.name	STRING	NULLABLE	Country name.
sessions.country.value	INTEGER	NULLABLE	Number of sessions.
sessions.source	RECORD	REPEATED	Sessions per source of referrals. For manual campaign tracking, it is the value of the utm_source campaign tracking parameter. For AdWords autotagging, it is google. If you use neither, it is the domain of the source (e.g., document.referrer) referring the users. It may also contain a port address. If users arrived without a referrer, its value is (direct)..
sessions.source.name	STRING	NULLABLE	Source name.
sessions.source.value	INTEGER	NULLABLE	Number of sessions.
release_date	DATE	REQUIRED	Last day of the release month. Table is partitioned on this column.