Work Log Day 5

Without a serious amount of new data, there was not a lot that could be done with the model that wouldn’t be suspect. As a way to pass the time I tried all of the different models that were defined here to see if any of them seemed significantly better. Sadly nothing jumped out as a real improvement over the base model.

The lack of data, particularly systematic data, is the real crux of the problem at the moment. I have decided to focus more on collecting for awhile. Up until now a lot of the shots are ‘ideal’, ie within the 30 second area, and I need to focus on collecting significantly more data outside of that range. My intention now is to collect 5 shots at a specific grind setting (seeing about a 5 second spread, likely due to tamping, distribution, etc) and see where the gaps in the data are. Once I identify the gaps, attempt to fill in the gaps.

The main progression of the day was that my 2lbs of Amazon Basics coffee came in, so spent the evening pulling shots in the 8-12 second range. Tomorrow will see more shots hopefully working up into the 30 second range.