The Value of Data

Last fall, I had my students construct and observe balanced aquaria. The project was only moderately successful. I had hoped to have the students work with the dataloggers, but the platform I was using proved to be a little too finicky. As with all student projects, the final data set was pretty messy:

But you can see some interesting trends and patterns. In two aquaria, pH increases over time, in the others it's flat. But across all of them, you can see a little uptick in pH when the light turns off, and then a return to baseline when the light comes back on. I'm not sure what that means -- it was not what I'd predicted (which was to see pH go up as CO2 was consumed the algal growth and, in the dark, for CO2 to accumulate and drive down pH. But, there it is: DATA! The world is more complicated than we think.

The technician who set up the incubator saw something quite different. What she saw was that the light was supposed to be off in the incubators for 8 hours and instead was only off for 2. They had been using these incubators for months with the assumption that the light was off for 8 hours and here was evidence that this was not so. Data!

After spring semester, we tested all of the incubators and found that they behave the same way. You program them kind of like a VCR (if anyone remembers what those are). After consulting with the company, she found that the programming doesn't persist across midnight, so you needed to set up two programs to have it be dark across midnight: one before midnight and one after midnight. Last night, we put the datalogger back and confirmed that the incubator actually turns off at 11pm and actually turns back on at 7am. Data!

I'm hoping when the summer moves on a bit, I'll have some time to actually develop a model for building data loggers that will make it easy for folks to implement them widely on campus. We have all the pieces -- I just need to organize them a bit so people can put them to work. Because more data is generally a good thing.