To understand hammerdirt it is necessary to recall the untidiness of the past. Karma is a bitch and you don't ever really know how big your debt is until it is paid in full. The last we heard from hammerdirt he was cleaning toilets thinking about his shiny new uniform. Things evolved from there, and things like that tend to have momentum. It takes a certain amount of effort to stop a large rock rolling down a hill (name your reference there).
Life goes on and datascience happened, hammerdirt tried to hang on for the ride and produced a django API that served a ReactJS client that was used to record observations in the field. These two bits of software were essential to produce a collection of JupyterNote books that combined made a pretty nifty JupyterBook. We added .pdf downloads for each chapter and put some serious effort into projecting the survey results onto a topographic map.
Reporting the results at an appropriate scale is one way of extracting value out of the data. This provides local stakeholders with an accurate presentation of the situation at an actionable scale. The next step in the value chain is to return those results back to the client/community in a way that they can use for their own research. The trash assistant is a GPT that holds the data for the region and a defined set of references.
The trash assistant is maintained by the people who collect the data and manage the repository. It is trained to perform certain standard tasks and has style preferences that can easily be modified. The maintainers ensure the integrity of the data and respond to issues concerning the trash assistant.
The trash assistant accompanies the static reports and can be used by client organisations to consider the results within the context their own projects. The trash assistant removes the need for