Data Used for Visualization in 2014
How do Differences in Growth Between Two Weed Species Affect Interactions with a Plant-Feeding Insect Used in their Biocontrol?
The study plants: Hawkweeds
The study insect: A plant-feeding gall wasp
Entomologist and agricultural research scientist, Dr Rose De Clerck-Floate (Agriculture and Agri-Food Canada, Lethbridge Research Centre), studies the intricacies of insect-plant interactions in order to use insects in the control of weeds. The plants being targeted for “biological control” are destructive pests of native grasslands, that have been introduced to North America from mainly Europe and Asia without the natural enemies that would normally keep them in check in their homelands. Weed biological control aims to reunite the most specific and effective of the insect enemies with their natural hosts. Some of the most specific and intimate of insect-plant interactions are between gall-forming insects and their host plants. Gall insects induce the production of unique plant growths, which protect and nourish developing, immature stages of the insect. The developing insects within the galls, however, also draw nutrients and resources from the plant that would normally go to plant growth and reproduction, thereby potentially curtailing the invasiveness of these plants.
For this exhibition, Rose provided the students with raw data from a greenhouse experiment that explored the interaction between a gall-forming, plant-feeding wasp (Aulacidea subterminalis) and the growing prostrate stems (also called “stolons”) of two host weed species of this biocontrol insect; “mouse ear” and “whiplash” hawkweeds. The experiment was set up to ask the questions:
· Is there a difference in how the two hawkweed species grow and develop as plants age?
· If yes, do species-based differences in growth and development affect plant-gall wasp interactions?
· What do the differences tell us about the potential efficacy of the gall wasp as a biocontrol agent of these weed species?
The experiment took into consideration the effects of both plant species (mouse ear and whiplash hawkweeds) and the stage of plant development at the time of attack by the wasp (“super young”, “young”, “mid”, and “old”). At harvest, the number of galls produced per plant and per stolon type (“main” stolons, which are the first to grow from the centre of young plants, and “lateral” stolons, which arise as branches from the main stolons later in plant development), which gives us an indication of what type of plant growth the gall wasps preferred or were tracking within plants. We also counted the total number of main and lateral stolons, which gave us an indication of; a) whether or not the galls had reduced plant growth, and/or b) whether the plant had responded to the presence of galls through stolon growth.
The students were initially given background information on the experiment, and the plants’ and insect’s biology, but were not told what answers Rose and her collaborating scientist colleague, Dr Rob Laird (University of Lethbridge, Department of Biological Sciences) found during statistical analyses of the data (see graphs). Thus what you will see in this exhibition represents the students’ individualistic artistic expressions of the data, but also their apparent empirical approaches to exploring and understanding the data and its revelations…….a process also generating new perspectives and creative lines of enquiry for the scientists.
- “Biocontrol” Simulation
The goal of this sketch was to make the data “invisible”, so that the viewer doesn’t even know that it’s being shown to them. Behind the scenes are many pieces of code that calculate the likelihood that a piece of the digital plant will be galled by the digital wasps. These calculations are based entirely on the data itself, taking into account the age of the plant, it’s species, and how many galls were created on it during the actual experiment.
This style of Data Visualization is an effective way of introducing the viewer to a set of data that is not overwhelming or intimidating, but also allows them to construct their own questions about what they are seeing, and (hopefully) seek to have them answered, either by careful observation, or by speaking with an expert in the field.
- Interactive Scatterplot
By Kiri Stolz
A scatterplot is a fairly basic form of data visualization. It is, however, incredibly legible and can be very easy on the eyes. In this example, the user is able to change what values are being represented on the X and Y axes, as well as the values being represented by the size of the squares themselves. The user is invited to explore the data, and make their own representation of it.
In the early stages of this sketch, I realized that there was more to my little scatterplot than I had originally thought. The spray of pixel-like squares on the display created a captivating pattern that I wanted to expand upon. Now, the user has the ability to create these abstract scattered designs that have a deeper layer of meaning to those who view it in the context of the data itself.
- Number of Galls on Main and Lateral Stolons
By Linda Shi
This model represents data collected by counting the number of galls found on main and lateral stolons of Whiplash. The four sizes of rings represent four stages of stolon growth from largest to smallest; we see the stolons in their old to super young stages. I have also chosen green zip ties to be the colour for main stolons whereas black ones represent lateral stolons.
By building this physical model, I’m able to see, group by group, on what type of stolon or which stage of stolon growth is the wasp more prone to lay their eggs. I’ve noticed that on the main stolon, wasps are more attracted to the young and super young stages of the stems, whereas on the lateral ones, more galls are found in the oldest stage of growth. One thing to keep in mind is that the lateral stolons sprout later than the main stolons. This difference in time could potentially result in the differences we see between galled stolons. I’m interested in the potential architectural installation of this model. In a built project, I want to see people become part of the data and have a 1:1 scale interaction with it.
- Cloud of Rings
By Linda Shi
In my final project, I’m taking the data I learned from Project one and the skills I’ve accumulated from Project two, to put together a larger skill installation in the presentation space.
My Cloud of Rings, instead of using the different sizes of rings to represent different stages of plant growth, I’m using the thickness of the rings to do so. By reversing the dark and light colors of the stolons, this color representation of the main and lateral stolons are more accurately displayed. Since the main stolons have grown for longer periods of time, they are thus represented by the darker colors.
Having the Cloud hang right above the hallway, I invite visitors to walk under the installation. The six strings hanging the are proportional to the number of total galls collected from each study.
The rings above overlap each other, creating various densities in the surface mesh. I hope this creates an interesting experience for those who walk below it. I hope to bring this project further by becoming more proficient at Rhino and Grasshopper. I would also like to use the 3D printer to construct smaller mock-up models of future interactions.
- Galls on Main and Lateral Stolons Voronoi Map
By Linda Shi
Taking the physical mock-up model one step further, my second project renders a pattern that evokes the same set of data from my first project.
Using 3D modelling software, Rhinoceros, and its plug-in parameter modifier, I was able to manipulate geometries on the plane. Using the parameter Voronoi, the maximum area is drawn around each point defined on the plane. I’m drawn to the mathematical and structural properties of the pattern, creating interesting dynamics between each of the defined points:
I want to use this pattern to represent the data collected. This pattern can easily be translated into a surface or screen that defines a specific architectural space. This will allow people to interact with data while navigating around the architectural installation.
By Kiri Stolz
This sketch represents the average number of galls in each of the six greenhouses that the bio-control experiments were carried out in. Each square represents one of the greenhouses, and the saturation of the colour is indicative of the average number of galls that had been created by the gall wasps at the end of the experiment. The idea here was to take the data back into the environment from which it came, and possible reveal new information about the space itself, and how that could have contributed to the production of galls or “fertility” of a particular greenhouse.
I was later made aware that some of the experiments were started later than others, which contributed to the low fertility rate in the top-leftmost greenhouse square. However, this data visualization led me to ask the questions necessary to learn more about the experiment than I had expected.
- Galled vs. Ungalled Portions of Hawkweeds
By Morgan Bath
All of the projects that I created this semester dealt with only the galled vs ungalled and did not focus on the lateral and the stolen figures. The works on paper are drawings of hawkweed seeds which were then manipulated to represent the ungalled portions of the plant that still have the ability to grow and the holes represent the lack of growth that the gulls created. The colours of the papers represent different plant growth stages as well as in the colour of the warp strings on the weavings I started.
The sounds piece that accompanies the works on paper is a creation from using the data provided for each plant stage and using the frequency of each number to create ten-second interval sounds of each plant stage and type. This piece I think was very successful, it is straight forward but I don’t think it is too literal, like the other pieces, right up front. I used two types of sounds wave one to represent galled and one ungalled and then just inserted the data set into the program to generate the sound clips.
- Re-Assessing Stolon Growth with Attention to Galling in Hawkweeds
By Keith Morgan
My previous visualizations had given me a few assumptions about the trial that seemed to go against some of the hypotheses of the researchers. I had come across a trend that seemed to show that when introduced to the biocontrols at certain ages, the hawkweeds tended to grow more stolons rather than fewer. What I had not taken into account when exploring this was what the galls themselves were doing: how were the galls dispersed on the plants, and did this have an effect? Again, I severely averaged out the data, this time combining all of the trials and all of the ages for the test group as one entity, and the control group as another. The key metrics for this visualization were the total average number of main and lateral stolons that grew on one plant, the average number of main and lateral stolons that had galls, and finally the average total number of galls per plant. Assuming that a galled stolon is considered ‘dysfunctional’, these new numbers actually reversed my original assumption that while galled plants tended to have more stolons on average, they actually had fewer ‘functional’ stolons. In my visualization, only these functional stolons grow to full length.
- Stolon Counts Between the Control and Test Trials of Whiplash and Mouse Ear Hawkweeds in Response to Galling
By Keith Morgan
Having worked with the data more, I was curious as to how much of an effect the galling process had on the quantity of stolons that grew on each species at the different trial ages. To achieve this, I compared the test group against a set of control samples as well. One of the primary questions for the original experiment was to test if the biocontrol agents could prove to be a natural method to control the hawkweed’s spread.
With this visualization, I maintained the age distinctions as it became clear when going through the data that the different species were more affected at different ages. I instead distilled the lateral and main stolon metrics together in order to have an average number of total stolons per species per age for both the control and test trials. With these numbers, I found the ratio of galled plants to the control plants in terms of the number of stolons they had. This ratio – while potentially misconstrued due to the averaging process – showed that only the Very Young group in Whiplash and Very Young and Young groups in Mouse Ear had reduced stolon counts. It appeared in the other groups that galling had in fact stimulated stolon production. It is this difference that I chose to express in my visualization.
A number of metrics are represented in this sculpture. The monoliths are 16’ high, 9.6” deep, 6’ wide, and have 48 spines. This represents the 16 plants per replicate, the 96 total plants in the trial, the 6 different replicates conducted, and the 48 plants per species used in the trial respectively.
The tree structures in this project were an addition inspired by Linda Shi’s first visualization where she modeled a series of abstracted trees using the data points to constrain the height of the trees and leaf population. When she explained that her model represented a large-scale installation that someone could walk under, my mind was completely and utterly blown wide open. It was the first data visualization that I resonated strongly with, and her architectural approach inspired me to create my own large-scale installation piece. Running within the Blender Game Engine, I wanted to honour Linda’s piece by recreating it in a setting more suited to her original vision alongside my own visualization.
- Comparing Gall Formation on Main and Lateral Stolons in Whiplash and Mouse Ear Hawkweeds
By Keith Morgan
This was my first encounter with the data set, so in order to get familiar with the metrics and patterns in the data, I chose a straightforward representation to keep myself grounded. I attempted to show the number of main and lateral stolons that had been galled, in addition to the total number of galls that had formed on each species. These sets were expressed as the height of each collection of blocks in relation to the top of the clear walls on the sculpture.
My process involved combining the data from all of the ages of the plants in each trial so that the comparison was left to differences between species instead of age. What this allowed me to do was show that proportionally speaking, the Whiplash species had more main stolons galled, while Mouse Ear had more lateral stolons galled. Still, when we look at the total number of galls that grew, it became apparent that both species grew an almost identical number of galls.
- Visualizing the Proportion of Galled and Ungalled Stolons Interactively
For this visualization, I looked at the relationship between different data points, particularly the proportions of stolons that were actually galled (which is the idea result of the experiment). I originally was working with the number of individual galls that were on each stolon, but realized that without seeing the relation of how many stolons were actually galled, the real message would be lost.
There is a lot of data being represented, and I really want people to be able to interact with it in order to create their own stories and get immersed in the data. There are two preset positions to help encourage these explorations.
This visualization calculates both the average number of galled and ungalled stolons per age group in each of the Hawkweed species, as well as the percentage that is actually galled. The lateral stolons really show how little proportion of galled stolons there are, showing that the effect of the gall wasps on the largest amount of growth of the plant was minimal.
- Crocheting the Relationships Between Galled and Ungalled Stolons
By Taelynn Graham
This visualization uses the exact same data as my interactive digital visualization. I decided to represent the length of each section with crochet stitches, and they became quite lengthy, which shows how much spreading the Hawkweed plants actually do. I want people to be able to really understand the proportion of stolons being galled, and having a tangible object that you can hold and stretch out really shows this in a way that isn’t quite as easily understood by looking at a digital shape.
My intent was to keep the colours the same for both visualizations, so that viewers an easily move from one visualization to the other and be able to recognize that these are in fact the same pieces of data. The mounted crochet strips with a single strand down them represent the stolons galled and the average number of galls on these stolons. This is an added piece of data included to show how many galls exist in each section, and what it means in comparison to the proportions of galled stolons. Just because there are a lot of galls does not mean that the proportion is substantial.
- Galled vs. Ungalled 3D Bar Graphs
By Corwin Smith
For this representation I wanted to show a comparison between galled and ungalled stolons in Whiplash. The green graphs represent the main stolons of Whiplash, and the yellow represent the lateral stolons of Whiplash. The graphs are separated into rooms, and then ages of the plants within the room. I always find 3D graphs on the computer to be very bad because of perspective, so I wanted to see if that same problem comes through when 3D printing the graphs of data. They were luckily much more effective as an actual 3D object than on the computer, and really allowed an interesting interaction from the viewer as they held it and were able to touch and look at it from all different angles.
- Stolon and Gall 3D Print
By Corwin Smith
This stolon and gall 3D print was made to show a larger scale representation of what the data was being collected from. Cylinders that fill the galls on the inside of the stolons give information about galled versus ungalled stolons relating to the age of the plant. This stolon shows the data collected from the main stolons of Mouse Ear. I was hoping to show what the stolon looked like with the galls to add to the understanding of what is being represented by this data. The cylinders used to represent area inside a gall were used to show the ratio between galled and ungalled in the different ages.
- Total Gall Count in the Main and Lateral Stolons for Mouse Ear Hawkweed
By Corwin Smith
The graphs represent the total gall count in the main and lateral stolons for Mouse Ear. The red represent lateral, yellow represents main, and green represents the total of both main and lateral. I was hoping to create these graphs to give an interactive aspect to them, where the user may swap them out and compare the different aspects of the data very easily.
Since it is a 3D object and can be moved around and interacted with, it brought a very unique feeling to the data when viewed the first time. It allowed myself to display a lot of data points efficiently and without the clutter that would happen with a similar 2d representation. This would allow them to compare parts of the data they need to.