After a whirlwind of evening training sessions and an excellent bit of training in the field with the GGI alums, the first week of the Fall DHS series went well! Here are the data:
Okay, there were a couple data hiccups, which took me a little longer than usual to troubleshoot. When I tried to run the analysis script in R, I got a series of unhelpful error messages – for some reason, the script was choking on the new data. It ran fine with data up to week 21, but there was an issue with the data in weeks 22 and 23. I figured out the first issues with some command-line sleuthing; there were 3 entries that entered time in the wrong format (e.g. “9.36” or 10.27″). I thought the data validation in the Google Sheet would have rejected those entries, but apparently it didn’t. The second issue took me longer to figure out, but it looks like some of the final measurements in week 22 may have actually been done on 9/4, which is technically the start of week 23. The script uses the date to determine the week, and got stuck when there were two entries in the same week. I bypassed that part of the script so that we could get the output, but you can see that the number of ALIVE plants in week 22 is too low, and the number counted in week 23 is too high:
We can still analyze these data, but have to careful about how we bin by week. You can also see that blip in the first panel below:
Data management issues aside, the other data should be more or less unaffected, and they show some interesting patterns. In particular, it looks like the plants are definitely starting to get smaller, and the larvae count might be starting to edge up over the last two weeks. Most importantly, it looks like a smooth transition between GGI and DHS – I don’t see any real discontinuities which would cause me to worry about data quality. I’ll be very interested to see what the data look like in the coming weeks – after a pretty intense heatwave, it feels like fall has finally arrived, with much cooler temperatures forecast for this week.