Setup
Here is pi-lot rigged up to a potted house plant. The bucket underneath has water that the pump can just about move up to a bottle on the plant’s trunk. The bottle contains the water sensor, and small holes in the lid. The pi is in a container hung on the outside of the pot. The moisture sensor and temperature probe are placed in the soil.
Apologies for the awful photography.
Plots
I’ve had pi-lot running for a little while on the plant. It’s been collecting data from its various sensors into an sqlite db. Now it’s time to have a look at the data.
Matplotlib is great and awful. Seaborn is prettier. Thought I might try bokeh for a more interactive visualization. I’ve used it before but briefly and it was a while ago.
EDIT : It would be cool to embed the figures but this is a lightweight static sight, and bokeh is not really in keeping with that.
Looks somewhat convincing.
Some correlation between the two temperature readings, and also with the humidity readings.
Day and night, and the evenings (when the lights are on) are picked up nicely by the photoresistor.
The water-level sensor hasn’t been particularly well utilized here … to put it kindly.
I’m not sure what there is to say about the moisture sensor.
So in summary: we can collect data!
Thoughts
My super niave implementation has left Bokeh struggling with a dataset that size. Something to think about in future.