torsdag 21 december 2017

Eating visually colossal insects (course)

I wrote a blog post about our new project course DM2799 "Advanced Project Course in Media Technology" in mid-November. No less than 7 out of 19 project groups worked with sustainability-related topics and I supervised three of these groups. All groups presented their results on Tuesday December 19.

In the course each group had to:
1) Conduct a small research project (defined by the faculty including our ph.d. students)
2) Write a project report (5 pages using the compact ACM paper format)
3) Present their project in front of an audience (15 minutes including a short questions-and-answer session)

I will briefly describe the results of "my" three groups here

1. Visualizing our colossal energy footprint (co-supervisor Mario Romero).
Three students (Andreas Almqvist, Per Jaakonantti and Alexandra Runhem) worked on "Visualizing our colossal energy footprint":

This paper discusses visualization in augmented reality (AR) in relation to the concept Homo Colossus. The aim was to explore how to visualize the concept of Homo Colossus in AR, and see if and what reflections the visualization could invoke about energy consumption. Using the game engine Unity3d in combination with the AR software Vuforia, an application that visualizes a Homo Colossus, based on a person’s annual flying habits, was created. The visualization was tested on 15 individuals on the KTH Royal Institute of Technology’s campus, aged 19- 27. It invoked deeper reflections among 8 out of 15 participants about energy consumption. Implications for information visualization in AR are discussed with a main focus on the importance of the user interface (UI) and the limitations in the used technologies

I have worked with the concepts "Energy Slaves" and "Homo Colossus" (in my a course on Sustainability and Media Technology). I find that these concepts help my students understand and visualize the (enormous) energy consumption in modern societeis. I have also started to write academic texts about these concepts, see for example this blog post about a submission to the academic track at the 75th World Science Fiction Convention (Helsinki August 2017) and this submission to another conference (that I in the end didn't attend). But why settle for writing about these concepts when we have great students who can help visualize these concepts with Virtual Reality (VR) and Augmented Reality (AR)?

This project represents a first attempt at using AR to visualize how large (in terms of calories, weight and height) humans would be if we were colossal creatures who ate as much energy as we use in our daily lives (for transportation, for heating our homes, manufacturing our gadgets, industrially agriculturing our food etc.). Instead of mapping the total amount of energy we use, these students settled for adding "only" the energy that is used for flying us hither and dither. They in fact used me as an example. I have flown four times in my job during 2017:
- to Brussels/Amsterdam to participate in an EU start-up meeting and to present a paper at a workshop (March 2017) - 4 hours of flight (275 kg CO2)
- to Denver to present papers at the CHI conference (May 2017) - 25 hours of flight (1300 kg CO2)
- to Amsterdam to teach at the ICT4S summer school (July/Aug 2017) - 4 hours of flight (275 kg CO2)
- to Frankfurt to participate in a workshop (October 2017) - 4 hours of flight (250 kg CO2)

My flight plans specify how large the carbon emissions are for these flights (see above), but I have also found this great online carbon emissions calculator for flights. I gave the students all the information (I forwarded my flight plans) but I now notice that they somehow added this up my figures and for some reason got 60 hours of flight instead of 37 hours. Or they might have just used a hypothetical professor - that is not me - who flew 60 hours in 2017.

Adding the energy used for flying me around for 60 hours to the energy in the food I eat, they ended up with my "shadow weight" being 8700 kilos, or "the weight of a tyrannosaurus rex" (or two killer whales). Instead of the 1700 kcal that I need to ingest every day, a 9-ton T-rex (or a 9-ton mammut etc.) needs to eat 63 000 kcal per day. Do note that the T-rex weighs ≈ 125 times more than I do but it "only" needs to ingest ≈ 37 times more energy because larger creatures are more "energy-efficient". 

The students also found information about how much the average KTH employee flies per year (6 hours). Adding these 6 hours of "shadow weight" to the (say) 70 kilos the average employee weights, they ended up with the average KTH employee weighing 380 kilos. This is equivalent to a fully grown moose.

The basic idea is to use the Homo Colossus concept is to communicate our rapacious energy consumption. My students help visualize it through an AR application on a modern smartphone where the blue human is a normal human and the bronze human is the colossal version of "me" (see image below). The size of the bronze human will depend on your 2017 flight habits, e.g. on the number of hours you have spent flying. Here's an example:

I have to point out that I haven't yet had time to check the students' numbers and calculations (which job categories are for example included in "the average KTH employee"?). I do however plan to continue to work with the Homo Colossus concept and will for sure define one or two new projects for next year's course.

The students had quite a few problems with the technology platform (Vuforia). One example is that you must point your phone at the marker (i.e. towards the floor, see image below). It would seem natural to change the angle of your phone and look up as the bronze human grows in size, but Vuforia won't work if your phone looses track of the marker on the floor so you will resolutely continue to look at the ankles of the giant bronze human... It also turned out that different phones (brands, models, operating systems) showed quite different things even though they of course are supposed to show one and the same thing.

Some ideas that was suggested for future work were 1) to morph human faces onto animals of various sizes, 2) to do a study that compares the understanding of regular charts vs AR visualizations and 3) to replace one huge human with numerous normal-sized humans.

2. Eating Insects would be good (co-supervisor Cecilia Katzeff).
Five students (Matilda Carlson, Lisa Lama, Oscar Linger, Gustav Mattsson and Raghu Nayyar) worked on the concepts of eating insects (instead of meat) and this resulted in the paper "Westerners' Willingness to Cook and Eat Insects: Reactions to a Prototype Website with Sustainable Recipes":

Insect eating provides environmental, nutritional and flavour benefits. However, western cultures have not adopted insect eating. This study used critical design to explore westerners’ willingness to cook and eat insects. A prototype website was developed that promoted sustainability and included insect recipes. Eight Swedish students interested in sustainability performed tasks in the website. The participants were told to motivate whether they would consider cooking and eating the dishes or not. The results show that the participants could consider cooking and eating various insect dishes. The analysis of the data showed that the willingness to cook and eat insect dishes depended on the 1) insect type, 2) insect form, 3) dish type, 4) dish presentation, 5) the context of cooking and eating, and 6) diets. The study shows that western students interested in sustainability could consider cooking and eating dishes from a sustainable recipe website.

This is a project group that have done just about everything right and they are currently my gold standard for what it is possible to accomplish in this short (7 weeks) but intensive (20 hours/week/student) project course. They specified their project clearly, were very smart when they recruited informants, did a good user study, made a great analysis of the results and wrote a terrific report. It's not quite good enough to be a "real" research paper but it's not too far from it either.

The question they studied was how open the environmentally conscious western student was to the idea of cooking and eating insects. They worked with "critical design" and "research through design" to examine users' assumptions about values and norms related to food and to idea of eating insects. They recruited 8 first-year students in the civil engineering programme "Energy and Environment Engineering" who studied the course "Ecology and Environmental Impacts". They then promptly misdirected their informants' attention by pretending to have designed a website with "sustainable recipes", but it "just so happened" that all the recipes they showed made use of insects. They displayed three types of recipes; first recipes with insect flour, then recipes with partially visible insect parts and finally fully visible insects (see images below). They also varied the type of insects (crickets, grasshoppers and mealworms) and, the images they used were spectacular. The report has so many interesting insights - like the comment that some vegetarians had problems when they tried to figure how they felt about eating insects ('ok' or 'not ok'?).

The students even made a point of challenging the audience when they presented their results. The pointed out that when they showed the first image in their slide deck (see image below), many people in the audience showed their discomfort and disgust by laughing nervously, but making faces or by covering their eyes. To show discomfort or disgust at the idea of eating insects is "normal" and nothing to be ashamed of - but we would frown if someone said that organic or vegetarian food was disgusting or display the same reaction to the suggestion that we should eat more vegetarian dishes. The conclusion is thus that we still have a long way to go before we get used to the idea of eating insects in Sweden.

Future work could be to not just discuss the idea of eating insects but to actually make people cook with insects and compare different groups. 

3. Affordances and limitations of ambient visualizations (co-supervisors Cecilia Katzeff and Mario Romero).
Four students (Emma Bäckström, Gustav Höglund, Tomás Albrecht and Victor Hedström Gustavsson) worked in a research project, "Sustainable Practices and data in urban areas – Opportunities for Change (SPOC)" and this resulted in the paper "Designing Ambient Visualizations to Encourage Organic Purchases: Using Research Through Design":

This paper examines how research through design can be used in order to design an ambient visualization to be used in a grocery store in order to encourage organic food purchases. The fundamental idea is that recurring customers would be able to see how the store performs in the aforementioned regard over time which, in turn, would encourage them to buy more organic food than before. The design process has been driven in an iterative manner where each design proposal has been thoroughly discussed and evaluated by using experts, user tests and brain storming to refine the visualization. The final iteration resulted in a visual structure representing a full year of organic purchases. To be able to see the real impact of the visualization further research and testing would need to be done. However, it turned out that it might be used for several other purposes than initially thought of.

The third and final group was special. They had already worked in our research project for half a year (but only for about 5 hours/week/student) and they were up and running long before the course started so they just continued their work and adapted it to the course and its format. Their main task has been to create a visualization that could help increase organic purchases in grocery stores and we have more specifically worked together with COOP in this project. The students have iteratively developed a number of prototypes of public visualizations that could be displayed in a store and they have struggled with how to best represent and present a lot of data in an attractive and easy-to-understand way. It would have been great to actually test the visualization for real, but they/the project ran out of time. The final version of the "dandelion" crams a lot of data into one image:

This image represents the collective purchases of organic foodstuff in a store. Each dot represents one day and each "spoke" represents one week. The greyed out dots on the top left represents the remainder of the year. The larger and more green the dot, the larger the proportion of organic purchases on that day and the smaller and more yellow the dot, the smaller the proportion of organic purchases on that particular day.

The most intellectually intriguing part of this visualization is the subtle manipulation that the students have introduced to make the image more attractive and easier to understand. Although it's hard to see, a "small yellow dot" is actually bigger if it is situated in the periphery compared to in the innermost circles. The spacing between the circles has also been subtly manipulated and these small "lies" maintain the "crowdedness" of the circles in the periphery compared to the core and they make the image a lot more interesting and easy to understand. With equal spacing, it is easy to get the impression (and it is just an impression) that the outermost circle represents a lower level of organic purchases due to the "spacing-out" effect with dots being further away from other dots. 

There is so much more to say about this work since we have worked with the students since mid-spring 2017. Below is for example an example of an early version of a dandelion but with a totally different mapping. The circle here represents a month (each "spoke" is a day) and the colorful circles represents various categories of food (dairy, meat, fish etc.).

All in all this was a fun course and my job as an advisor/supervisor was much reminiscent of being the supervisor of a bachelor's or a master's thesis. I was unfortunately totally unprepared for the fact that all my three proposals would attract students and for having to meet each group (at least) once per week throughout November and December - on top of a already heavy work load. It kind of worked out in the end, but my work load during this period has been exceptionally high and the plans for the Christmas break is to do as little as possible.

I have suggested that some of "my" 12 students (above) could continue to work on these topics in their master's theses during the spring, but the enthusiasm has been limited. The fact that I'm going to be on a sabbatical and won't be able to supervise any students at all doesn't help either, but I still think that the combo Advanced Project Course + Master's thesis opens up many possibilities for coming years. We should of course exert ourselves and create project course proposals that slide right in to master's thesis proposals and also prep students for continuing to work on their topics already when they start the project course in early November!

1 kommentar:

  1. Happy to be part of SPOC and see these student projects take form