Components & Compatibility

Components

Objectiv consists of the following components:

  • A taxonomy to give your datasets a generic & strict event structure designed for modeling.
  • Tracking SDKs for modern front-end frameworks to collect error-free user behavior data.
  • An open model hub with pre-built product analytics models & operations.
  • A modeling library to create reusable models that run on your full dataset.
What's in the box?
B.Y.O Data store

Objectiv connects to your own SQL cloud data store of choice. Objectiv Up (a pre-packaged, dockerized version of Objectiv for self-hosting) comes with a pre-installed Postgres data store.

Open analytics taxonomy

Objectiv is built around an open analytics taxonomy: a universal structure for analytics data that has been designed and tested with UIs and analytics use cases of over 50 companies. It ensures your dataset covers a wide range of common analytics use cases and is structured with modeling in mind. You can extend it to cover custom requirements as well.

The Open Analytics Taxonomy

Datasets that embrace the taxonomy are highly consistent. As a result, models built on one dataset can be deployed and run on another.

We're continuously expanding the coverage of the open analytics taxonomy. Support for marketing campaign analysis has been added recently, and areas like payments & CRM are on the roadmap.

Tracking SDKs

Supports front-end engineers to implement tracking instrumentation that embraces the open analytics taxonomy.

  • Provides validation and end-to-end testing tooling to set up error-free instrumentation.
  • Support for React, React Native, Angular & JS, and expanding.

Open model hub

A growing collection of pre-built product analytics models and functions. You can take and run them directly, or incorporate them into your own custom models.

  • Covers a wide range of use cases: from basic product analytics to predictive analysis with ML.
  • Works with any dataset that embraces the open analytics taxonomy.
  • New models & functions are added continuously.

Bach modeling library

A pandas-like modeling library to build models that run on the full SQL dataset.

  • Includes specific operations to easily work with datasets that embrace the open analytics taxonomy.
  • Pandas-compatible: use popular pandas ML libraries in your models.
  • Output models to production SQL directly, to simplify data debugging & delivery to BI tools, dbt, etc.

Compatibility

We aim to be compatible with the full stack of any modern data team. It is our goal to:

  • Support all modern front-end frameworks for data collection;
  • Support all major cloud data stores; and
  • Work or integrate with popular solutions for ELT, modeling, and output.
Compatible technologies & frameworks

Front-end

Objectiv’s Tracking SDKs are currently available for React, React Native, and Angular. A Browser SDK is also available to instrument Objectiv’s tracking without a specific UI kit. Our goal is to support all modern front-end frameworks with an SDK, including (but not limited to) Qwik, Vue/Nuxt, Next.js, etcetera.

Data storage & back-end

The open model hub and modeling library currently work with PostgreSQL and Google BigQuery, with Amazon Athena coming soon. Our goal is to support all major cloud data stores, including (but not limited to) Databricks, ClickHouse, etcetera. As a backend, Snowplow is supported as well.

If you want to self-host your own data store, you can use Postgres. To run Objectiv at scale, we recommend setting up a Snowplow pipeline, or use Objectiv Cloud.

Modeling & analysis

Objectiv’s modeling library and the open model hub can be used in any Python-based notebook (e.g. Jupyter Notebook, Google Colab, Hex and Deepnote). Check out the modeling docs for more information on how to get started in your notebook.

Output

While Objectiv’s models are built with a pandas-like interface, under the hood, they all run SQL directly on your data store. You can output any model to SQL with a single command, e.g to feed into other tools like dbt or BI dashboards. A tutorial on how to do that with Metabase (open-source BI) is coming soon.

Objectiv’s DataFrames can also be exported to pandas, which enables you to tap into pandas’ rich machine learning ecosystem (e.g. SKLearn).