Cube.js is an open-source, modular framework for building analytics into your web apps. You can use Cube.js for internal business intelligence tools, or customer-facing analytics.
Use this template to build your own open-source Google Analytics alternative. It's developer friendly, hackable, and embeddable. It is easy to install the full working application and then customize every part of it from data collection to and visualizations.
Artyom and Pavel from Cube.js here -
We've been receiving a lot of questions and requests from the community on how to build a web analytics platform, similar to Google Analytics, with Cube.js. That led us to building a ready-to-use application and template for this, which everyone can easily install and customize.
By building and hosting such an application yourself, you'll be able to fully control the privacy and don't need to share your users’ data with 3rd parties. It also means no vendor lock-in and you completely own the underlying data.
We've designed this template to be fully hackable and embeddable. This means it's very easy to customize on every level—from data collection to metrics definitions and visualizations. The frontend is a pure React application based on Material UI without any custom styles. You can embed any part of the frontend into your existing application and customize the look and feel to match your styles.
Last, but not least, we've focused a lot on the performance to keep response time under 50 ms and make it scale up to several million daily active users.
We hope you enjoy it and build a lot of great and useful applications with this template. Happy hacking!
@keydunov That sounds really cool. It should be much more flexible then g-cloud and mixpanel.
A couple of questions:
1. does it support regular stuff like cohort analysis?
2. I run my own DB, right? Any support for custom DB queries?
3. If custom queries are supported, can I go a bit further and score data against my ML models? How am I supposed to serve the model?
@fullstackml Thank you! Since it's not a monolithic application, but rather a template designed for customization on any level it is easy to add all these custom things.
1. The template itself doesn't have it, but it is very easy to add with Cube.js data schema, which powers metrics definitions.
2. You can customize all the queries and add new ones but changing Cube.js data schema definitions.
3. We don't go into ML space directly, but I believe it should be possible to create data schema against the tables in the database populated with the results of ML models.
@nikitakorotaev Thank you! It is a great question. Unlike Matomo, it is not a monolithic application, but rather a set of modules, which based on different established open-source technologies: Snowplow, PrestoDB, Cube.js, and React with Material UI on frontend. You can change or
replace any component with technology that better fits your needs. Additionally, due to Cube.js data schema you can customize how metrics and dimensions are defined in a case the default definitions don't work for you.
Cube.js
DVC
Cube.js
Cube.js
Cube.js