Introducing Pushbroom Analytics
Last updated
Pushbroom Analytics is a project that was born out of asking basic questions around how human beings are using a website or tool, informed by ethical data collection practices and rigorous data science. Pushbroom looks to answer the basic questions; What’s happening, and how do we know?
There are a lot of really excellent, privacy-first analytics tools available today — many of them from small, independent developers. These tools are really good, and have set the industry standard for what ethical site usage analysis can look like. They include, incompletely;
Pushbroom wouldn’t exist without these great tools and the vision set by their developers. Any one of these tools can probably the job you need, but there are a couple of things that they don’t do which led me to develop Pushbroom instead of paying for a service and being done with it.
Pushbroom is an event-based tool that collects zero data by default. Events with sessions in a way that many of these tools don’t, creating space for insight into common session paths. Every event in Pushbroom is also associated with its previous view url, allowing for path analysis of user flows. This creates a rich view of how a site’s content is being experienced.
Pushbroom can accept arbitrary data, and treats all event key:value pairs as first class, queryable data points. These events can be ultra-wide, deeply nested, and capture complex relationships between sessions, views, and other events.
Pushbroom uses W3C standards for open, interoperable data. All of the data in Pushbroom can be exported and transformed into any context it might be needed. Additionally, it’s not a requirement that Pushbroom store any of your data. If you have a standards-implementing data store, Pushbroom rely on that instead. This means your data is yours, and you can take it anywhere you want.
All of this means that the stage is set for robust data analysis to answer questions about how your site is being used. What events are statistically correlated with what page views? What content is being viewed and what is not? What content views are associated with sessions that perform later events? Any question we can ask of our usage patterns can be answered — without any invasive or privacy-violating data being collected.