Mixpanel dbt Package (Docs)

What does this dbt package do?

Produces modeled tables that leverage Mixpanel data from Fivetran's connector. It uses the Mixpanel event table in the format described by this ERD.
Enables you to better understand user activity and retention through your event data. It:
- Creates both a daily and monthly timeline of each type of event, complete with metrics about user activity, retention, resurrection, and churn
- Aggregates events into unique user sessions, complete with metrics about event frequency and any relevant fields from the session's first event
- Provides a macro to easily create an event funnel
- De-duplicates events according to best practices from Mixpanel
- Pivots out custom event properties from JSONs into an enriched events table

Generates a comprehensive data dictionary of your source and modeled Mixpanel data through the dbt docs site. The following table provides a detailed list of all tables materialized within this package by default.

TIP: See more details about these tables in the package's dbt docs site.

Table	Description
mixpanel__event	Each record represents a de-duplicated Mixpanel event. This includes the default event properties collected by Mixpanel, along with any declared custom columns and event-specific properties.
mixpanel__daily_events	Each record represents a day's activity for a type of event, as reflected in user metrics. These include the number of new, repeat, and returning/resurrecting users, as well as trailing 7-day and 28-day unique users.
mixpanel__monthly_events	Each record represents a month of activity for a type of event, as reflected in user metrics. These include the number of new, repeat, returning/resurrecting, and churned users, as well as the total active monthly users (regardless of event type).
mixpanel__sessions	Each record represents a unique user session, including metrics reflecting the frequency and type of actions taken during the session and any relevant fields from the session's first event.

How do I use the dbt package?

Step 1: Prerequisites

To use this dbt package, you must have the following:

At least one Fivetran Mixpanel connector syncing data into your destination.
A BigQuery, Snowflake, Redshift, PostgreSQL, or Databricks destination.

Databricks dispatch configuration

If you are using a Databricks destination with this package, you must add the following (or a variation of the following) dispatch configuration within your dbt_project.yml. This is required in order for the package to accurately search for macros within the dbt-labs/spark_utils then the dbt-labs/dbt_utils packages respectively.

dispatch:
  - macro_namespace: dbt_utils
    search_order: ['spark_utils', 'dbt_utils']

Database Incremental Strategies

Many of the models in this package are materialized incrementally, so we have configured our models to work with the different strategies available to each supported warehouse.

For BigQuery and Databricks All Purpose Cluster runtime destinations, we have chosen insert_overwrite as the default strategy, which benefits from the partitioning capability.

For Databricks SQL Warehouse destinations, models are materialized as tables without support for incremental runs.

For Snowflake, Redshift, and Postgres databases, we have chosen delete+insert as the default strategy.

Regardless of strategy, we recommend that users periodically run a --full-refresh to ensure a high level of data quality.

Step 2: Install the package

Include the following mixpanel package version in your packages.yml file:

TIP: Check dbt Hub for the latest installation instructions or read the dbt docs for more information on installing packages.

packages:
  - package: fivetran/mixpanel
    version: [">=0.10.0", "<0.11.0"] # we recommend using ranges to capture non-breaking changes automatically

Step 3: Define database and schema variables

By default, this package runs using your destination and the mixpanel schema. If this is not where your Mixpanel data is (for example, if your Mixpanel schema is named mixpanel_fivetran), add the following configuration to your root dbt_project.yml file:

vars:
    mixpanel_database: your_database_name
    mixpanel_schema: your_schema_name

(Optional) Step 4: Additional configurations

Collapse/expand details

Macros

analyze_funnel (source)

You can use the analyze_funnel(event_funnel, group_by_column, conversion_criteria) macro to produce a funnel between a given list of event types.

It returns the following:

The number of events and users at each step
The overall user and event conversion % between the top of the funnel and each step
The relative user and event conversion % between subsequent steps

Note: The relative order of the steps is determined by their event volume, not the order in which they are input.

The macro takes the following as arguments:

event_funnel: List of event types (not case sensitive).
- Example: '['play_song', 'stop_song', 'exit']
group_by_column: (Optional) A column by which you want to segment the funnel (this macro pulls data from the mixpanel__event model). The default value is None.
- Example: group_by_column = 'country_code'.
conversion_criteria: (Optional) A WHERE clause that will be applied when selecting from mixpanel__event.
- Example: To limit all events in the funnel to the United States, you'd provide conversion_criteria = 'country_code = "US"'. To limit the events to only song play events to the US, you'd input conversion_criteria = 'country_code = "US"' OR event_type != 'play_song'.

Pivoting Out Event Properties

By default, this package selects the default columns collected by Mixpanel. However, you likely have custom properties or columns that you'd like to include in the mixpanel__event model.

If there are properties in the mixpanel.event.properties JSON blob that you'd like to pivot out into columns, add the following variable to your dbt_project.yml file:

vars:
  mixpanel:
    event_properties_to_pivot: ['the', 'list', 'of', 'property', 'fields'] # Note: this is case-SENSITIVE and must match the casing of the property as it appears in the JSON

Passthrough Columns

Additionally, this package includes all standard source EVENT columns defined in the staging_columns macro. You can add more columns using our passthrough column variables. These variables allow the passthrough fields to be aliased (alias) and casted (transform_sql) if desired, although it is not required. Data type casting is configured via a SQL snippet within the transform_sql key. You may add the desired SQL snippet while omitting the as field_name part of the casting statement - this will be dealt with by the alias attribute - and your custom passthrough fields will be casted accordingly.

Use the following format for declaring the respective passthrough variables:

vars:
  mixpanel:
    event_custom_columns:
      - name:           "property_field_id"
        alias:          "new_name_for_this_field_id"
        transform_sql:  "cast(property_field_id as int64)"
      - name:           "this_other_field"
        transform_sql:  "cast(this_other_field as string)"

Sessions Event Frequency Limit

The event_frequencies field within the mixpanel__sessions model reports all event types and the frequency of those events as a JSON blob via a string aggregation. For some users there can be thousands of different event types that take place. For Redshift and Postgres warehouses there currently exists a limit for string aggregations (up to 65,535). As a result, in order for Redshift and Postgres users to still leverage the event_frequencies field, an artificial limit is applied to this field of 1,000. If you would like to adjust this limit, you may do so by modifying the below variable in your project configuration.

vars:
  mixpanel:
    mixpanel__event_frequency_limit: 500 ## Default is 1000

Event Date Range

Because of the typical volume of event data, you may want to limit this package's models to work with a recent date range of your Mixpanel data (however, note that all final models are materialized as incremental tables).

By default, the package looks at all events since January 1, 2010. To change this start date, add the following variable to your dbt_project.yml file:

vars:
  mixpanel:
    date_range_start: 'yyyy-mm-dd'

Note: This date range will not affect the number_of_new_users column in the mixpanel__daily_events or mixpanel__monthly_events models. This metric will be true new users.

Global Event Filters

In addition to limiting the date range, you may want to employ other filters to remove noise from your event data.

To apply a global filter to events (and therefore all models in this package), add the following variable to your dbt_project.yml file. It will be applied as a WHERE clause when selecting from the source table, mixpanel.event.

vars:
  mixpanel:
    # Ex: removing internal user
    global_event_filter: 'distinct_id != "1234abcd"'

Session Configurations

Session Inactivity Timeout

This package sessionizes events based on the periods of inactivity between a user's events on a device. By default, the package will denote a new session once the period between events surpasses 30 minutes.

To change this timeout value, add the following variable to your dbt_project.yml file:

vars:
  mixpanel:
    sessionization_inactivity: number_of_minutes # ex: 60

Session Pass-Through Columns

By default, the mixpanel__sessions model will contain the following columns from mixpanel__event:

people_id: The ID of the user
device_id: The ID of the device they used in this session
event_frequencies: A JSON of the frequency of each event_type in the session

To pass through any additional columns from the events table to mixpanel__sessions, add the following variable to your dbt_project.yml file. The value of each field will be pulled from the first event of the session.

vars:
  mixpanel:
    session_passthrough_columns: ['the', 'list', 'of', 'column', 'names']

Session Event Criteria

In addition to any global event filters, you may want to disclude events or place filters on them in order to qualify for sessionization.

To apply any filters to the events in the sessions model, add the following variable to your dbt_project.yml file. It will be applied as a WHERE clause when selecting from mixpanel__event.

vars:
  mixpanel:

    # ex: limit sessions to include only these kinds of events
    session_event_criteria: 'event_type in ("play_song", "stop_song", "create_playlist")'

Lookback Window

Events can sometimes arrive late. For example, events triggered on a mobile device that is offline will be sent to Mixpanel once the device reconnects to wifi or a cell network. Since many of the models in this package are incremental, by default we look back 7 days to ensure late arrivals are captured while avoiding requiring a full refresh. To change the default lookback window, add the following variable to your dbt_project.yml file:

vars:
  mixpanel:
    lookback_window: number_of_days # default is 7

Changing the Build Schema

By default this package will build the Mixpanel staging models within a schema titled (<target_schema> + _stg_mixpanel) and Mixpanel final models within a schema titled (<target_schema> + mixpanel) in your target database. If this is not where you would like your modeled Mixpanel data to be written to, add the following configuration to your dbt_project.yml file:

models:
    mixpanel:
      +schema: my_new_schema_name # leave blank for just the target_schema
      staging:
        +schema: my_new_schema_name # leave blank for just the target_schema

Change the source table references

If an individual source table has a different name than the package expects, add the table name as it appears in your destination to the respective variable:

IMPORTANT: See this project's dbt_project.yml variable declarations to see the expected names.

vars:
    mixpanel_<default_source_table_name>_identifier: your_table_name

Event De-Duplication Logic

Events are considered duplicates and consolidated by the package if they contain the same:

insert_id (used for de-deuplication internally by Mixpanel)
people_id (originally named distinct_id)
type of event
calendar date of occurrence (event timestamps are set in the timezone the Mixpanel project is configured to)

This is performed in line with Mixpanel's internal de-duplication process, in which events are de-duped at the end of each day. This means that if an event was triggered during an offline session at 11:59 PM and resent when the user came online at 12:01 AM, these records would not be de-duplicated. This is the case in both Mixpanel and the Mixpanel dbt package.

(Optional) Step 5: Orchestrate your models with Fivetran Transformations for dbt Core™

Expand for details

Fivetran offers the ability for you to orchestrate your dbt project through Fivetran Transformations for dbt Core™. Learn how to set up your project for orchestration through Fivetran in our Transformations for dbt Core setup guides.

Does this package have dependencies?

This dbt package is dependent on the following dbt packages. These dependencies are installed by default within this package. For more information on the following packages, refer to the dbt hub site.

IMPORTANT: If you have any of these dependent packages in your own packages.yml file, we highly recommend that you remove them from your root packages.yml to avoid package version conflicts.

packages:
    - package: fivetran/fivetran_utils
      version: [">=0.4.0", "<0.5.0"]

    - package: dbt-labs/dbt_utils
      version: [">=1.0.0", "<2.0.0"]

How is this package maintained and can I contribute?

Package Maintenance

The Fivetran team maintaining this package only maintains the latest version of the package. We highly recommend you stay consistent with the latest version of the package and refer to the CHANGELOG and release notes for more information on changes across versions.

Contributions

A small team of analytics engineers at Fivetran develops these dbt packages. However, the packages are made better by community contributions.

We highly encourage and welcome contributions to this package. Check out this dbt Discourse article on the best workflow for contributing to a package.

Are there any resources available?

If you have questions or want to reach out for help, see the GitHub Issue section to find the right avenue of support for you.
If you would like to provide feedback to the dbt package team at Fivetran or would like to request a new dbt package, fill out our Feedback Form.

bdtoole/dbt_mixpanel