CROP
CROP copied to clipboard
CROP is a Research Observation Platform
There are plans in place to use the Hyper.ag platform to control (initially) heaters, and (later) ventilation units. They would use temperature measurements from individual sensors to determine when to...
Jakob says that having the crop data quite up to date would be very useful. Sub 5min should be sufficient, although even faster wouldn't hurt. Could also think about something...
As a remnant of the massive amount of unused code and dependencies we used to have in the front-end, `custom.js` and `custom.css` (under `webapp/app/base/static/`) have a lot of code that...
Zack has been investigating [Dynamic Mode Decomposition (DMD)](https://towardsdatascience.com/dynamic-mode-decomposition-for-multivariate-time-series-forecasting-415d30086b4b) as a potential method to predict conditions inside the farm. This method could also take into account the spatial nature of the...
The R postgresql interface we are using apparently has an issue where requesting too many rows of data causes datetime information to be truncated to just dates. This seems to...
We have David joining our team from the start of September. I think this would be an excellent time for us to update our documentation, mainly on the project level,...
There are a number of improvements to the GES code we would like to make, but before embarking on those, the code could use a general updating of documentation and...
Turns out, we have a wiki! It's got, among other things, setup instructions for how to get started: https://github.com/alan-turing-institute/CROP/wiki/1.-Set-up-and-deploy-CROP-locally-or-on-Azure It's all quite out of date. I've removed the link to...
The ARIMA model is currently running on a desktop workstation. It should be deployed using an Azure Function.
If we calibrate the GES model using VPD instead of RH then we get worse matching of RH and T, but better matching of VPD. So if VPD is the...