Day 2

Today was another busy day, but I have no complaints as I really enjoyed the workshops. First, we were given a presentation on IRB and Ethics. Here we learned about the 3 Main Principles of Human Subject Research and the different types of approval needed from the IRB when completing certain studies.

After the presentation, we were taught a basic overview of how to use Tableau. It took me a second to get the hang of it, but I was amazed at how quickly you could change a filled Excel spreadsheet into a nice, pretty and easy to read visualization. I thought it was actually really cool. Prof. Siek then gave a short lecture on qualitative shadowing and observing, which my roommates and I were able to test out at Starbucks tonight. I used the method called object mapping as I recorded the use of cellphones among the customers at Starbucks. Sadly, my computer crashed when I got back, deleting my notepad file I had left open filled with my observation notes!!!!! *angry face* My fault for not saving the file, but I am pretty mad about it. Although I do not remember specific times, I will do my best to highlight what I noticed specifically during my observation:

  • A woman next to me was on her iPad, and used it as a surface to prop up her phone which played a video tutorial of how to draw something. In front of the ipad and phone was a sketch pad, where she practiced drawing from watching the video. Although watching the video, every 2 minutes or so she would readjust her phone so that she could what looked to me like responding to a text message. After some time, she switched to a movie on her ipad and moved her phone to right next to her sketch pad so that she could respond to texts in a faster manner.
  • A man sitting in front of me with long shaggy blond hair seemed to be working on a paper of some sort. Surprisingly he was one of the few that did not have a phone out – but not for long as he pulled out his phone from his book bag and left the store completely (phone and coffee in hand)…but he left his stuff to go outside for a smoke break.
  • A man and woman sit together at a table, conversing. Yet every few minutes one of them pulls a phone out of their pocket and halts conversation to check their phone.
  • Another man  multiple tables in front of me sits with a textbook in front of him on the table and a laptop placed in front of the book. Despite the obvious work he probably needed to get done, he is sitting with his hands under the table holding his phone. Interestingly enough he received a call and went outside to take it (it was pretty quiet in Starbucks at this hour of the evening). When he came back inside though, he once again ignored the open book and laptop and continued to mess with his phone, which was now held in his hands almost on top of the laptop keyboard. He appears to be scrolling a lot, which leads me to think he is on some kind of social media application or perhaps looking at the news. Why was he on his phone and appearing to look something up when he could have simply used the open laptop in front of him??
  • A man walks into the store with a small bag in one hand and a large smart phone in the other. He walks up to the counter with his phone screen lit and eyes towards it rather than the employee standing at the register. He didn’t even turn off his phone while ordering, let alone made any sort of eye contact with the employee. Hmm, that was kind of rude.
  • Another woman walks in clutching her phone. There’s no one standing in line, and the store is kind of empty and quiet. She didn’t have to wait to order or anything, but she takes out her phone and starts scrolling and typing away as an employee had to walk over to her (but was still close!) Once again she barely makes any eye contact with the employee and keeps her phone out and in use while she waits for her name to be called.
  • A group of three teenage girls walked in complaining of how quiet it was. Two out of the three were clutching their cell phones, and one was even using it in her hands (head facing the ground) as she opened and walked through the entrance. While waiting for her drink they all take out their phones, but the girl who was on hers the whole time’s drink was ready and it took a second to realize her name had even been called because she was so absorbed in her phone! How hard is it to miss your name being called in a quiet lonely Starbucks??
  • Finally, I observed a man sitting in a chair reading a book. He catches me staring at him a few times which is kind of awkward because I think he knows that me and my roommates are all observing him…he puts down the book and takes out a phone. Types away then continues to read. Then, he suddenly pulls out another phone (what? two phones???) and scrolls away on that one as well.

I’d also like to add that during this 20 minute period, Olivia and Vanessa were communicating over text message on their phones sending texts to me and Devon about how the guy reading in the chair totally knew that we were all watching him…awkward. Even more interesting is that while sitting INSIDE STARBUCKS Olivia used her phone to order it instead of going up and standing in line. *Ah, the power of technology* However, if this was a real study I was conducting, when looking into the IRB and Ethics behind it, i think it would be exempt as it is an “observation of public behavior,” even though that guy probably felt a little weirded out by us…

Tonight I also read a paper entitled Relational Learning for Sustainable Health. Here, I put together a summary of the main points. I found this article to be very interesting and would definitely like to keep tabs on this research as I do agree that the use of predictive models could ultimately develop a more personalized health journey for individuals (especially those who cannot afford expensive treatments, hospital visits, etc).


In this paper, the authors claim that in order to manage a sustainable society, methods that can predict the occurrence of critical health-related events are crucial when trying to develop personalized health plans for society. Through three separate case studies that addressed important problems including predicting cardiovascular complications years in advance, predicting the population that needs to be monitored for developing Alzheimer’s in their later years, and predicting adverse effects of drugs based on the individual’s characteristics and their medication regimen, the studies’ results indicate that the potential of the algorithms used to develop predictive models could possibly lead to realizing the main idea of  personalized medicine.

In the first case study, experiments revealed that the relational learning methods are more suited for the task of predicting later cardiac events by analyzing the young adulthood data due to their ability to handle multi-relational data. Also, risk factors from early childhood (of the subjects) are seen in the experiments to be the most important factors when it comes to indicating risks at later years. This allows populations to take control of their cardiovascular health earlier and develop treatment plans before confronted with the risks of cardiovascular events later. Socioeconomic risk factors appear to be as predictive as clinical risk factors (bmi, bp, etc), which opens up another avenue for potentially developing sustainable health care for cardiovascular risks. In the second case study, those who were considered to be in a population that had increased chances of developing Alzheimer’s disease later in their lives needed to be predicted. In order to reduce the extreme amount of costs it takes to care for someone with Alzheimer’s, it is crucial to identify those who are at risk for Alzheimer’s earlier in life in order to prevent future problems. The first step in the process of being diagnosed with Alzheimer’s disease is isolating those who appear to have MCI (mild cognitive impairment). Results showed that relational models were quite effective when it came to isolating those who have MCI. Estimating the number of those who have MCI that will actually go on to have Alzheimer’s would develop a sustainable society where costs are balanced between treatment for individual patients and society. This goal can be reached by developing patient centric Alzheimer’s treatment plans.

In the third case study, it was found that learning in reverse is a good alternative for unsupervised learning problems when the labels are not known in advance as in the case of adverse drug events. What is important is the censoring of the data based on the learning task, as it is important to omit data about patients before they started the drug. Also given the multi-relational nature of the data, it was necessary for algorithms that directly operate on relational data without flattening it. THUS relational methods are a natural choice for modeling tasks that are essential in developing a sustainable society – as identifying potential side-effects of drugs in the first years after their release to a large population has the possibility of reducing individuals cost for treatment, reducing hospital admissions and preventing catastrophic events.

So as the initial results of these three case studies demonstrate the potential of the algorithms used to develop predictive models, in the future, authors suggest that it is important to develop these algorithms to handle multiple modalities of data so they can extend to handle subsets of the data types in order to make effective predictions. Machine learning algorithms usually are applied to individual modalities, but should be extended as results obtained from one study or group in a population have to be validated across the data of other groups in order to understand the data fully and draw inferences. Basically, by employing machine learning to develop predictive models for a person’s health, an overall idea of personalized medicine will ultimately lead to more sustainable health care practices.

Also, today we discussed ShareLatex (which was a relief because I know it!) and Github. Github was sadly not as easy as I do not have an extensive knowledge of coding and all of its vocabulary. Basically what I got from playing around with GitHub is that it is the equivalent of GoogleDocs for code. Hopefully I get to use it more to be more comfortable with it.

The Faculty Topic Talks really interested me today – I was particularly interested in Patrick Shih’s study with online health communities as I have read lots of work on rare diseases and understood the concepts he was lecturing about. Also, being a part of a few online health communities myself, I relate to those in them and the lack of support they often receive in real life when it comes to their chronic illness.

To practice using ShareLatex, we were asked to create two different types of short bios: one being professional, and one that would interest those in K-12. Here I include a PDF of my ShareLatex document:


I feel that I understood ShareLatex well but still need to work on using references.

Well, it’s looking like it will be another long night as I have more work to complete to get ready for Laser cutting and 3D printing tomorrow Here’s a sneak peek at a design I’m going to 3D print for my little brother at home. Hope he likes it!