Tuesday, November 27, 2007

Garden Ideas

Here is a brief synopsis of the ideas and development i've done for the visualization component of the Technology Garden, compiled for 11/29 meeting.

Charlotte and I had discussed my creating an interactive visualization (and sonification) of the sensor data being collected from the garden. The idea is for there to be a wall-mounted touchscreen display that hangs over the plants and complements the current informational display with something more representational and ambient. This display will allow for different ways to visualize the current and past state of the garden; encourage interaction with the digital portion of the garden's ecology; become another facet of the aesthetic experience of the garden; and can act to communicate garden state externally - within the building and via the internet.

I got some sample data from Bryant that they collected during the pilot study. Here is a simple bar chart visualization of the readings over time. I'm just parsing the data file and doing a little bit of normalization for the display.

There are 812 samples in the dataset for (5) soil moisture sensors and temperature and humidity sensors for the room. Bryant said it was roughly 1.5 days of data with 20 seconds between each sample. However with 812 samples it seems like it is more like a sample every couple of minutes (or the data covers a shorter period of time). This first reading of the data it appears that the room temperature and humidity have an inverse relationship to one another... which makes sense i think. yes? The climate fluctuations are likely between daytime and nighttime settings of the climate control. With the soil moisture, a higher reading means dryer soil. I believe this dataset starts with the sensors not in the soil (to explain the really high readings) and then we see a watering and a slow taper after the watering.

Next step I think we need to capture some more data and do these things:
  • Timestamp the samples so we can get a more accurate sense of time
  • Figure out how to normalize and calibrate the readings between the different moisture sensors as best as possible
  • Begin archiving sensor data in a database. Will be a tradeoff between sample frequency and storage space.
  • Begin simple visualization of realtime data so we can get a better sense of how the data changes in real time.

Other sensor data that we would like to capture:
  • Light readings, both inside and outside the building. Outside readings will reflect the sun patterns and help bring an outdoor context. Indoor light readings might help infer activity in the room. The interaction between these two light sources can add interest to the visualizations.
  • Motion sensor or webcam to discern movement and activity in the room. Could be one sensor that just reflects presence in the room, or with more work could reflect movement in different zones, or reflect when people enter vs. leave the room.
  • Microphone to gauge sound. Do not want to record the actual audio (no surveillance!) but just to record sound levels to also help infer when there is activity in the room. For instance, if there is motion in room but no sound then someone is present but not talking. Sustained sound level might reflect social activity of some sort.
  • Silvia and I had talked about trying to measure the vibrations of the plants (she was interested in touch between the people and plants) and thought about using piezoelectric for this. Any kind of microscale vibrations of the plants would be cool. Also have read something about microvoltage fluctuations in plants but need to do more research.
  • Oh a new one... how about seismic? I know Calit2 has seismic sensors we could probably tap into and perhaps so does Bren Hall? I think it is interesting to consider the way that earth and building vibrations might relate to the vibes of our little micro ecology :)
Phase One (real time display)
Display the sensor information in real time through visuals and sound. The sensor acquisition computer will relay the readings via a network socket to the client application running on the touchscreen display. The default screen will give an overview of the state of the garden (plant moisture, light reading, motion, sound). Interaction with the display allows for more detailed information on a plant-by-plant or sensor-by-sensor basis. "Events" can also be reflected through sounds, graphics or messaging on the display. For instance, we can discern a watering event by the sudden change in moisture level of a plant sensor... the plant representation can reflect this watering and perhaps a sound or a message can be delivered. (also see below for ideas on how to use this event to collect further data from the people in the garden) Other events could be someone entering the room or turning on/off the lights... perhaps a welcoming message or a change of state of the visualization.

Integration of historic sensor data. Allow for interaction with the display to provide the browsing of historic data through timeline-based visualizations. Also (as above) allow the visualization of historic data for a specific plant or sensor. View animations depicting the garden state over the course of a day, week, month, etc. When were plants added, "friends" made, and other events can be included.

The display could become an interesting part of the social component of the project. It can reflect current social relationships that exist in the system... such as plant sponsorship, friends, etc. As we build more social features into the system we could use the display to facilitate. For instance, we have talked about wanting to display garden visits and maybe tying events to people. While we may want to be careful about making the system too surveillant, we could collect voluntary information. For instance, a watering event could thank the user and ask them to identify themselves. We then have some more social information to display and to develop activity patterns and linkages between plants and people.

People. Plants. Technology. Some people are sure to ask "why?" Many of the benefits of the garden exist simply between the plants and the people... relaxation, becoming more familiar with natural or "sustainable" ideas and systems, a place to gather, etc. So what is technology's role in this, does it make the system more sustainable? The CHI paper shows how technology can be used to help engage gardeners and get them involved, with the garden and with each other. Technology brings the garden to the gardeners when they can't be there in person, and helps to bring them back to the garden when they can come. In many ways the technology acts as a link between the people and the garden. In some ways this could be construed as making the system more sustainable, but in general I think we really need to unpack the implications of this term, especially if we are using it as the center point of the rhetoric around the project. If increasing awareness of sustainability is a goal we maybe should figure out ways to emphasize this in the project in a more overt way.

A Hybrid Ecology
Another way to look at the relationship between plants/people/tech that I'm interested in pursuing with the visualizations is the notion of a digital ecology that exists in a symbiosis with the other two components - the human ecology and the garden ecology. Inside the visualization I would like to enable an ecology of digital organisms that have behavior attributes which cause them to interact with the garden in various ways. Some organism may like light, others vibrations, some attracted to high moisture while others are repelled. Maybe even some parasitic organisms or some kind of food chain interactions. Gardeners program simple character attributes and release their creatures into the wild of the garden, where their fate depends on interaction with plants and environmental factors and on interaction with other digital organisms (think a simple version of Spore with "real world" influences from a biological system). I think this has potential to be a very cool thing... theoretically and as far as being fun. One possibility is that when someone waters a plant they are presented with the award of being able to add a new digital creature to the garden. This will allow the digital ecology to grow over time with the garden, provide a personal connectedness to the digimals (digital animals... sorry just made that up), and could increase awareness of the ideas behind enabling a sustainable ecosystem. I have many more ideas surrounding this, but this blog post is already way too long!