ga-attribution-scrape 0.0.10

Scrapes attribution data from GA through JS Network in Python for CSV exports.

ga-attribution-scrape

Scrapes attribution data from GA through JS Network in Python for CSV exports.

Notes

• The program assumed separate conversions and does not currently try to sum conversions together from separate conversion IDs.
• When dealing with GA goals, will pull all goals as separate requests.
• Works on a Service Account for authentication.

How to run

First import the Scrape function:

from ga_attribution_scrape import Scrape

Then initialise ga_attribution_scrape with the Scrape() function which must contain a config dictionary, which can be found at https://github.com/lewisaustinbryan/ga-attribution-scrape/blob/main/empty_config.yaml

config

Service Account

You have to create a service account in Google Cloud Platform that has Bigquery access and GA access if you want to use a goal as a kpi for attribution reports. Help on creating one can be found here. https://cloud.google.com/iam/docs/creating-managing-service-accounts Separate Service accounts can be created for GA and Bigquery

There are four main parts to the configuration:

GA

Here you need to include account ID, Property ID and view ID.

Bigquery

For including the dataset ID and Table ID to tell Bigquery where to put the attribution reports.

Backdate

If backdate is True then will just pull yesterdays data, otherwise it will loop through each day on the specified start_date and end_date

Unless you explicitly set GOOGLE_APPLICATION_CREDENTIALS in the environment (e.g. using os module), be aware that the program expects you to backdate first with service account, then when backdate is False it refreshes the service used. There is no other option but it makes it very easy to put into a Cloud/Gamma Function.

Request

This is where we copy the request from the JS network in Google analytics in the "Conversions -> Multi Channel Funnels -> Model Comparison Tool" report for the request url https://analytics.google.com/analytics/web/exportReport/, which will have various query parameters associated with it.

Copy everything from Request Headers and and Form Data, which is included in the empty_config.

Get Attribution Report

`# Import get_ga_goals: from ga_attribution_scrape import get_ga_goals

Get the goals from GA management API and create a pandas dataframe:

ga_goal_management_data = get_ga_goals(accountId, propertyId, viewId, credentials) # where credentials is service account #Now Scrape the GA attribution report for each goal according to configuration: Scrape(config).goals(ga_goal_management_data)`