PyAmplitude

...What is Pyamplitude ?

Amplitude its a python-based client designed to interact with the entire set of Amplitude Analytics services (Redshift, Export API, Dashboard Rest API and Behavioral Cohorts API) following the official Amplitude Documentation which can be found at:

https://amplitude.zendesk.com/hc/en-us/categories/115000290488-Technical-Resources

General Organization:

 AmplitudeRestApi   -> A wrapper around each Rest API resources.
 RedshfitAmplitude  -> A sql connector for the Amplitude AWS-Redshift database.
 ExportApi          -> A wrapper around the entire Amplitude export

Aditional Modules:

Resources      -> Api resources ( Manage several apps & secret keys, import Segments to get data using Segment definitions )

Aditional Features:

- Calculate each query cost.
- Manage your query rate limits in a ETL context.
- Visualize the quentity of requests and your query limit status.

Install

pip install pyamplitude

How to use PyAmplitude ?

Let's start by importing the ProjectHandler and passing a project_name, api_key and api_secret key as parameters.

from pyamplitude.projectshandler import ProjectsHandler

Lets instance the ProjectHanlder with our first app called 'BubbleWay'...

bubbleConector = ProjectsHandler(project_name='BubbleWay',
				 api_key='<api_key>',
				 secret_key='<secret_key>')

Hint: Use the repr method to check your actual instance when used with several apps.

print bubbleConector

project_name: BubbleWay | api_key: API-KEY | secret_key: API-SECRET

from pyamplitude.amplituderestapi import AmplitudeRestApi

Querying data from the Amplitude Analytics REST API Dashboard

Great ! So let's use the Amplitude Rest Api to query some useful data for later analysis...

apiconector = AmplitudeRestApi(project_handler = bubbleConector,
                               show_logs       = False,
                               show_query_cost = False)

Creating and using segments

Segments are represented as JSON arrays, where each element is a JSON object corresponding to a filter condition. First import and initialize the Segment class and add each query filter. in these cases we will be creating two segment definitions and for later use.

first_segment_definition = Segment()

first_segment_definition.add_filter(prop='country',op='is',values=['argentina','brasil'])

second_segment_definition = Segment()

second_segment_definition.add_filter(prop='country',op='is',values=['argentina','paraguay'])

Using the AmplitudeRestApi module.

First Example: Querying Active and New Users count

data = apiconector.get_active_and_new_user_count(start    = '20170814',
                                                 end      = '20170825',
                                                 m        = 'active',
                                                 interval = 1,
                                                 segment_definitions = [first_segment_definition,
                                                                        second_segment_definition],
                                                 group_by            = None)

If all goes well... you should receive a JSON response such as:

{
    "data": {
        "series": [
                    [18600,15294,14164,12945,12585,11797,10113,9523,8321,7873,9053,8109],
                    [3264,3423,3397,2984,2916,2827,2918,2934,1800,1560,1240,1100]
        ],
        "seriesMeta": ["Argentina", "Brasil"],
        "xValues": ["2017-08-14", "2017-08-15", "2017-08-16", "2017-08-17", "2017-08-18", "2017-08-19", "2017-08-20",
                    "2017-08-21", "2017-08-22", "2017-08-23", "2017-08-24", "2017-08-25"]
    }
}

Other Resources: Session Length Distribution, Average Session Length, Average Sessions per User, User Composition, Events, Events List, Event Segmentation, Funnel Analysis, Retention Analysis, User Activity, User Search, Real-time Active Users, Revenue Analysis, Revenue LTV, Annotations...

Aditional options

Using calculate_query_cost = True parameter

With PyAmplitude you can calculate each query cost very easily by checking the show_query_cost parameter. In an ETL context you may use this option not to exceed query limits. If you want to know more about how each query cost is being calculated, please read:

https://amplitude.zendesk.com/hc/en-us/articles/205469748-Dashboard-Rest-API-Export-Amplitude-Dashboard-Data#request-limits

Fetching data from Amplitude Redshift

As a addition you can query data from Amplitude Redshift using the AmplitudeReshift module.

from pyamplitude.amplituderestshift import AmplitudeRedshift

ConnectionHandler = AmplitudeRedshift(host='',
                                      user='BubbleUser',
                                      port='5439',
                                      password=<yourpassword>,
                                      dbname='bubble_db',
                                      schema= 'app123098',
                                      table=<db_table>,
                                      show_logs= True)

query = 'SELECT * FROM app123098.bubble_db LIMIT 10'

ConnectionHandler.execute_query(query=query)

NOTE: AmplitudeRedshift has a serious of prebuild methods to fetch a list of users, query user events as well as new users, but you can also pass your specific query using the execute_query method.

Authors

Marcos Manuel Muraro