I want to take this opportunity to praise SnowPlow.
They allow for collection from just about anywhere. Options include Javascript, GIF beacons for email embed, Arduino, and anything else that can send HTTP GET requests.
Snowplow's collection can be performed by Amazon CloudFront, which results in absurdly low latency world-wide, for cheap. I think I estimated around $0.60 per million hits. However, Snowplow is very decoupled, so collection can also occur on NodeJS servers, or anything else that conforms to their spec. In fact, everything is fairly neatly decoupled, so if you think something would run better in [FAVORITE_LANUAGE/PLATFORM] you are free to implement it.
I run Snowplow on my blog, but only run collection for right now. Eventually, when I get around to setting up the rest of Snowplow, I'll be able to crunch the numbers. This is huge - most analytics services require processing and collection to be closely tied.
All that said, Snowplow isn't plug and play like Google Analytics or even Piwik is. You need to be willing to patiently wade through the Snowplow documentation.
Both Piwik and Snowplow emulate Google Analytics too closely. Using the same syntax is silly, and limits their potential. I'm still waiting for an entirely JSON based analytics service.
On the JSON/GA point - thanks for the feedback. We are slowly but surely evolving from our GA-style API, e.g. with our tracker support for new Mixpanel-style unstructured events (though they are not - yet - supported in our ETL/storage):
We are moving our event model to be Avro-based later this year and then yes we may look at completing our existing tracker protocol with something more JSON-centric. So lots on the horizon!
Comments
I want to take this opportunity to praise SnowPlow.
They allow for collection from just about anywhere. Options include Javascript, GIF beacons for email embed, Arduino, and anything else that can send HTTP GET requests.
Snowplow's collection can be performed by Amazon CloudFront, which results in absurdly low latency world-wide, for cheap. I think I estimated around $0.60 per million hits. However, Snowplow is very decoupled, so collection can also occur on NodeJS servers, or anything else that conforms to their spec. In fact, everything is fairly neatly decoupled, so if you think something would run better in [FAVORITE_LANUAGE/PLATFORM] you are free to implement it.
I run Snowplow on my blog, but only run collection for right now. Eventually, when I get around to setting up the rest of Snowplow, I'll be able to crunch the numbers. This is huge - most analytics services require processing and collection to be closely tied.
All that said, Snowplow isn't plug and play like Google Analytics or even Piwik is. You need to be willing to patiently wade through the Snowplow documentation.
Both Piwik and Snowplow emulate Google Analytics too closely. Using the same syntax is silly, and limits their potential. I'm still waiting for an entirely JSON based analytics service.
Thanks so much for the words of encouragement!
On the JSON/GA point - thanks for the feedback. We are slowly but surely evolving from our GA-style API, e.g. with our tracker support for new Mixpanel-style unstructured events (though they are not - yet - supported in our ETL/storage):
https://github.com/snowplow/snowplow/wiki/2-Specific-event-t...
We are moving our event model to be Avro-based later this year and then yes we may look at completing our existing tracker protocol with something more JSON-centric. So lots on the horizon!