A few observations:
- The Sea Leveler is driven by twitter’s own search API, which is not perfect. The rapid dramatic drops are due to twitter updating its search parameters to exclude tweets more than a week old. Thus, the Sea Leveler records increased activity in real time and decreased activity less frequently, but in larger steps.
- I didn’t line the paper up very well, so the dates and times aren’t perfectly calibrated.
- The massive drop on 9 April is slightly more than a week after Boing Boing picked up the Sea Leveler, thus reflecting the tweets resulting from the coverage being purged from the search.
- The vertical lines that drop and return quickly are errors in the search function. The Sea Leveler is programmed not to move if the search function returns 0 tweets (which would indicate a connectivity problem).
- I don’t know what was going on on 12 April, but @johnvanderhoef and @lindsaycthomas were tweeting up a storm from what sounds like a very interesting series of talks.
I’ve currently set the Sea Leveler to record a full month on one roll, so be sure to check back in May to see how it’s going. Until then, keep talking about #sealevelrise!
Two weeks ago, I announced my latest Hacking the Ocean project, an open-source, Arduino-powered water level meter that monitors the frequency of tweets containing the hashtag #sealevelrise. Since launch, the Sea Leveler has had some bugs and received some good press. Now that I’ve had some time to monitor its performance and work the bugs out of its code, it is finally time for the promised “how to build the Sea Leveler” post.
This project was much more involved than my Arduino build and significantly more rewarding. The Sea Leveler was a challenge on multiple fronts, from learning to make the Arduino talk to twitter to physically modifying the water level meter. As I noted in my first project log, I have very little programming experience, and the major goal of this build was to level up my C++ skills. I’m very happy with the results, both technical and aesthetic.
For simplicity, I’m going to break this into two posts, one for hardware and one for software.
Earlier this week I launched the Sea Leveler, and open-source, arduino-powered, water level meter that measure the activity of tweets about #sealevelrise on twitter. Not surprisingly, the first full week of trial revealed a few bugs in the machine.
The first thing you’ll notice is that, in addition to recording tweets about sea level rise, the Sea Leveler also provides a nice documentation of power surges. Every time the power flickers, the arduino resets and the arm thinks it’s back at zero, causing a dramatic rise. This happened once due to an actual power surge and twice due to our marvelous dishwasher, which happened to be on the same circuit as the Sea Leveler. Easy fixes, both.
The second problem is that, thanks to our very cheap step motor (you get what you pay for) after a few power cycles, the unit starts rotating in the opposite direction. Wonderful. Obviously the permanent fix is to get a slightly higher quality motor, but for now, isolating the circuit to reduce power surges will have to do.