infryq

Yesterday I met with V about freebase preprocessing, ran support for the nowcasting team to help get their signal ready, finished a program to generate a sample of data for L in an Excel-friendly format, and ran an experiment for R to see if expanding our definition of the community question for the F data helped at all (it did not). At the indicators meeting A presented a fix for the issue where standard errors become zero when the estimate is zero, and N had evidence to show that our best guess for why the G and F community questions are on such different scales is wrong, so merging them will not be as easy as we had hoped. Gave a brief update at the team leads meeting. Dinner, crossword machine was briefly broken so I cleaned up a few signal renaming tasks that came in late (why did it take this long to identify suboptimal naming? grr) and watched DS9, then crossword, snack, and sleep.

I finished Rogue Protocol and started Exit Strategy. These are better the second time around but that may just be pandemic stress.