Scoop me!

Care to make a contribution? Email UCSDScoopMe@gmail.com and I'll update this site with new suggestions and any kudos I come across over time.

ScoopMe! arose from many conversations over the years with the punchline "that's a good idea, but I'm not going to be the one to do it." One of those ideas, years back, was a site to organize all the un-done ideas, in hopes that those looking for inspiration (students, entrepreneurs, hobbyists, etc) could make better use of the ideas than if the ideas just disappear back into the mist. This is not that site - making that is implicitly on this list - but it's a start.

• How different is the life of an urban vs rural squirrel? Squirrels are a lot like people - diurnal, social, handsy, intelligent, and opportunistic. So how different is life for those in cities vs. the wild. Are the commuter squirrels? Do city squirrels have diabetes and breed year-round because light pollution obscures seasonal signals? Are rural squirrels getting more or less zoonoses, or more or less steps per day? Possible project for tiny sensors with antenas in trees and on garbage cans.

• Is good writing differentiable as a fractal? Many distributions in nature are fractal, and fractal distributions often have an aesthetic quality to them. If one takes all work by an author, and looks at sentence length (words or characters), is that a (roughly) fractal distribution? If so, or to the extent it is, does that fractal dimension identify an author's works differently than, for example, word usage frequency?

• Transforming digital story telling with baby boomers. Every culture has rituals around listening to the elders tell stories, because (in part) this is a way to download experience vicariously to the next generations. The baby boomers will be both the largest and the most documented cohort for how human life ends that has occurred. How do we expand our story telling rituals beyond the verbal (or film) into the realm of data-driven health and socio-behavioral analyses and models? What kind of legacy can an aging individual leave behind through their data to inform the health journeys of those that follow, and how can that be safely shared and curated to support sustainable social wellness across diverse groups of people?

• How do off-the-shelf classifiers fair on different kinds of biomedical data? Even pre-college students can use off-the-shelf (prepackaged) machine learning code to build classifiers. Places like PhysioNet provide many examples of different kinds of biomedical related data, from EKG to EMR samples. With a systematic program, one could have students run the same "most common" classifier packages over every data set in PhysioNet (done systematically, not all at once by one persons). Scores could be compared to provide an empirically-mapped landscape of which kinds of classifier seem to most consistently work on which types of biomedical data. This would enable deeper, mechanistic explorations of why the trade-offs observed arise, but could also help guide younger students to more fruitful first research efforts.

• Can smart collars detect cat cancer risk? A common cause and result of many diseases, including diabetes and cancer, is disruption of daily rhythms. Can pet ailments be identified early with pet wearables, like smart collars? If so, what are the implications for pet health, differences by pedigree, and implications for human medicine?

• Are cats a net health gain? Cats have been shown to lower cardiovascular disease risk by reducing stress through contact, petting, etc. However, they can also increase sleep disruption and so may contribute to other long-term health risks. A community research project collecting biometrics, wearable data, and or self-report from cat owners could help to create a real-life map of cat impacts on health for different demographics. Combined with the "smart collar" concept, one might even imagine identifying "felotypes" that best predict different levels of sleep disruption, stress reduction, etc.

• What are the real academic fields? Academics, scientists, and scholars of all stripes tend to find differences from others' work to identify themselves. But not all specialties are equally different. One way to map similarities would be to analyze word usage (in UCSD faculty web sites, in journals, etc.) and create network models to answer questions like: how many clusters are there, really? How distant or even distinguishable are different fields and subfields? Are there elemental dimensions along which fields fall, like more or less quantified, more or less social, or is there more going on than we intuit?

• How biodiverse is your block? How many species of grass, flower, tree, herb, insect, mammal, bird, amphibian can you find if you look every day? Over time, these observations could be compared to street lighting, noise, traffic, water quality, on and on...

• Do spiders have chronotypes? When are different spiders building webs, out and about, or inactive? Could make a good K-12 pooled observations effort.

• How sparse can readings be to accurately model changes in coupled oscillator networks? Given some range of couplings across oscillators, and some range of perturbations of interest at different oscillating nodes, model the frequency of sampling necessary to usually catch the perturbations. Useful in determining mini-max for devices and medical tissue collection.

• Can we detect latent racism in traffic patterns? Sleep deprivation increases likelihood of traffic accidents. Shift work and working multiple jobs affects sleepiness, and is the labor often disproportionately represented by minorities and immigrant communities. Can we see more late night traffic accidents in these communities, and if so,  can interventions be developed with these communities.

• How much "noise" in molecular biology is really just unmeasured rhythmicity? The microbiome, metabolome, proteome, etc., are all part of a body system that oscillates around the day (circadian), within the day (ultradian), and across many days (infradian). These component/output omics should show signatures from these rhythms, presumably with some rate-determining limit in resolution. Across individuals these rhythms tend to wash out, but what about with repeat, rapid sampling within individuals? What percent of variance could be shown to correspond to physiological rhythms?

• Can smart phones make light exposure maps? Given that smartphones are often continuously sampling the visual environment, could these observations be aggregated to create a living light-exposure map of the world at night?

• How does verbal complexity inform selection in song birds? If sound is providing information from the environment, then does growing up in a more complex soundscape promote selection for added complexity of song (and sound/song recognition capabilities) in song birds? The idea being that the birds need to be recognized among greater clutter, but also parse that clutter.

• Are plants and animals parallel vs serial solutions? Plants and animals need to seek out sources of energy and water. Plants (largely) do this with parallel branching growth for leaves and roots, and then regress unsuccessful branches. Animal (largely) do this by serially exploring spatial pathways, and revisiting more rewarding pathways. Is there an abstract search algorithm that can be used to test the relative efficacy and limits of the parallel vs serial path solution to living?

• Can you build a death alert system? As a person approaches death in a hospital setting, there are many physical signs that practitioners recognize. Loss of muscle strength appears to be among the most reliable. It seems like a wearable sensor could be devised to detect probable impending death, which could be used to give families time to join the passing, and perhaps enable interventions before such final signs as heart failure.

• What's your idea?

• Want to do it as a project? Email me and we'll discuss a project.

• Want someone else to do it? Email me and we'll post it.