As we welcome 2017, it’s time to sum up the key developments in data science and machine learning from the past year so that we can open our eyes to the new year. Here is what you missed …
Source: Data Science and Machine Learning Developments in 2016 – Radiology Data Quest
Posted in Figure Stuff Out
In an MVP-based approach, flagship design begins with a paper boat. And a mouse.
The concept of MVP applies to research, quality improvement, or innovation projects. In this case, MVP is not “most valuable player” (although it could make you one) but “minimum viable product.” In a nutshell, rather than the traditional approach of building only after deliberate planning and careful design, the MVP concept focuses on a just-enough set of features. At first glance, it may seem counter-intuitive: shouldn’t the greatest success come after long-term project planning and thorough discussion of all possible outcomes?
Initially, MVP was developed for anemic startup companies to get a quick infusion of revenue before developing their flagship offering. It was soon realized that this process of small-target rapid iteration yields not just faster but also better results. Gantt charts, project timelines, and table-pounding meetings are still important, but real-life experimentation is a higher priority.
MVP is an extension of Plan-Do-Study-Act (PDSA). It makes one assumption: “all assumptions are more likely wrong than right.” In designing a medical research proposal, you have implicitly made assumptions about some aspect of a disease’s biology. In creating a product, you inevitably face the need to make assumptions about customer needs. In starting a business, you have determined that about research outcome, or even the fundamental design.
If these assumptions are more likely wrong than right, then the best next step must be to make as few as possible before having the opportunity to test them. The MVP approach asks innovators to encapsulate as few concepts as possible into a deliverable and then bring that product to a test market to verify those assumptions. Since outcome metrics can be very difficult or expensive to obtain – just ask people who run Phase III trials or market research – limiting variables allows you to be sure that acquired data only have a small number of possible interpretations.
Some approaches to software design approaches embrace the MVP concept, one of the better known being Scrum. Another product-oriented approach is called pretotyping (not to be confused with prototyping) – “faking” as much as possible with the goal of acquiring data before the making heavy financial investments.
The venerable Harvard Business Review has more – there are two types of MVPs. Your MVP can be validating – by trying an inferior product to prove a concept. It can also be invalidating, where the MVP is actually a better project than the one you plan to create.
If your MVP is a worse product than your imagined final version, success validates your idea; failure, on the other hand, doesn’t necessarily invalidate it. If your MVP offers a better experience, then failure invalidates your business model; success doesn’t necessarily validate it.
Hypothetically, if your radiology department is contemplating an investment of $3 million over 5 years on a “virtual radiology consultation” technology to improve communication, the rationale for purchase may be that busy radiologists cannot satisfy the high clinician demand for collegial discussions, and live digital discussions would solve that problem.
To test this assumption, you could deploy an invalidating MVP. For instance, this may take the form of a one-week real radiology consultation for all questions.
This solution is obviously a huge waste of valuable radiology resource and unsustainable over time. But failure invalidates one key assumption for the intended purchase. Even if successful, it may raise important points to resolve: is subspecialist availability necessary? Does the consultation need to be 24/7 or only during key hours? The virtual solution might still work, but it would work for reasons other than but you know at least one of the underlying assumptions might need reassessment.
Posted in Figure Stuff Out
A gifted young baseball player played Little League through college ball, hitting an average of nine out of 10 pitches. [But] hitting .900 in the Major League is impossible. The best baseball players […] miss two out of three swings. They sometimes strike out in a crucial moment, drop balls, make bad plays, and disappoint their fans.
The difference […] is that world-class baseball players wake up every game day ready to swing the bat.
Source: John Donahoe: Dump the Myth of the High Achiever | Stanford Graduate School of Business
The RSNA conference will take place in Chicago in 1 month. If you’ve already started looking at the meeting program, you might get the same sense of excitement as the rest of us – this …
Posted in Figure Stuff Out
After the recent presidential election, you are probably either particularly alarmed or especially excited about the outcome. Regardless of your particular political predilection, it is fair to say that this election puts data science on its head when so many got so wrong.
On an earlier issue of Harvard Business Review, the venerable magazine shared a piece of research from the University of Southern California on forecasting. When forecasting sales, the best estimators use a combination of intuition and logic – with both the logic-heavy and intuition-heavy forecasters performing less accurately.
In the age of artificial intelligence and big data, it can be sobering to realize that despite the staggering volume of data we are now collecting, ignoring your gut instincts can take a heavy toll on your decision-making abilities.
Source: “What type of forecaster are you?” Harvard Business Review (March): 26.
Posted in Figure Stuff Out, Technology and Informatics
Tagged Data science, Predictive Analytics
The first part of this discussed the heterogeneity of data projects and how a uniform approach can help hone in the solution. The first post also discussed the first two elements: Refine the question, and identifying the right data. Here we tackle the next two elements.
At this step, we begin to go into the technicalities of data science. This post is not designed to go into the detail of each approach, but it will attempt to ask the relevant questions.
How will you process the data that you now possess? In almost all cases, this step will involve data wrangling (also known as data munging or data cleaning). To determine how the “clean” form your data must take for proper analysis, it is important to determine the transformations and algorithms necessary for your question. Continue reading
My wife and I take a routine monthly trip to Costco to refill the refrigerator. Now with less than two months from core exam, she said she can drive by herself so I can have more time to study. It was thoughtful of her to offer. I thought for a moment. Buying chicken and cheese may be routine and unexciting, but it is something we do together, and there are some things more important than doing well on a test.
Posted in Figure Stuff Out
“The result of our approach,is that we end up with a team of people who will quickly become bored by performing tasks by hand and have the skill set necessary to write software to replace their previously manual work.”
Google engineers are not afraid of automating themselves out of a job. They embrace the challenge of finding the next best thing in machine learning, in big data, in medicine, or moonshots like longevity, because of this philosophy.
Are we bored with clicking and measuring things by hand yet? Spell checking your report manually for semantic (i.e. error of meaning not spelling) errors? Making a differential diagnosis strictly from memory? We should get bored. Then we can start to improve it.
It’s when we are satisfied from “good enough” that we forget “doing better” is possible.
… and imagine if you could program life itself. Rather than 0’s and 1’s, you have four possibilities, a computing system performing quaternary arithmetics.
I still remember being dazzled as a freshman in college, during the first computer science lecture. The professor spoke of quantum computers, where improvements in speed of calculations can be measured in squaring time 2n rather than the traditional doubling time (i.e. Moore’s law) 2n. And there was biologic computing, using simple building blocks of genetic material ACTG to perform calculations which take place in living cells.
Then, I spent the 15 years that follows writing them off as science fiction, pontifications of an old man.
I was, of course, wrong.
When I was in 8th grade, my English teacher wanted to give everyone a book to take into high school. She had a cardboard box full of various books. There was literary fiction like Toni Morrison. There was a memory aid for American presidents. But I came to class really late that day, so by the time I went up to the box, there were only a few books left. I had the great choice between Billy Budd (dryest. book. ever.), Atlas Shrugged, and this book called Getting Things Done.
I picked up Getting Things Done because Atlas Shrugged didn’t fit in my bookbag. It would be years before I realized that self-help productivity books is in itself a major genre of nonfiction. At the time it just didn’t make sense why anyone would need such pathologic level of compulsion to keep things organized.
