This post is part of a series of posts and testimonies featuring data science students from the Data Science Immersive 12 Weeks, Full-Time Career Accelerator in Washington, DC offered by the educational company General Assembly. Students share what they've learned, what they wish to accomplish and what they are doing next.
Yoni Levine graduated from the Data Science Immersive Class of June 2017. Shout! News is presenting his current research, given the broad appeal and articulate presentation of his analysis on using technology to help classify suicidal risk in social media. Read on!
September 11, 2017 • Yoni Levine
Yoni's interview follows:
Q&A with General Assembly's Data Science Immersive Class of June 2017 graduate, Yoni Levine, and Flavius Mihaies, founder of Shout! News.
Shout!: Congratulations on your graduation from the Data Science Immersive at General Assembly! I was wondering what prompted your decision to pick this topic that Shout! News profiled? Is that something you had in mind for some time already?
Yoni Levine: Healthcare in general, and specifically mental health are of particular interest to me. I spent several years working with people who had both physical and developmental disabilities, and found tremendous meaning in what I did. Before I decided to take the Data Science Immersive course I was about to begin nursing school and decided, instead, to couple my interests and use my data skills in the health industry. There are so many exciting things happening in the field now, so when I got to choose my own project for my capstone at General Assembly I knew that I wanted to work on something healthcare related.
I spent several years working with people who had both physical and developmental disabilities, and found tremendous meaning in what I did.
Shout!: What was the main challenge, editorially, data wise, or both that you encountered and how did you address it?
Yoni Levine: Finding the perfect data-set is always really hard, and finding healthcare data is especially so. There are so many laws protecting the data, as well as subject matter research concerns to be aware of that it is almost impossible to get your hands on any interesting medical data. I had to get creative in my search, and I decided to scrape a public forum where people seeked support for depressive and suicidal feelings, and analyze that.
Shout!: On a different note, data-journalism is really at the intersections of two fields that did not much talk to each other until recently. Do you see this as a challenge? How do you picture the future of data-journalism? Is this a career attractive to young data science graduates these days?
Yoni Levine: Writing is something that I love to do, and data journalism allows me to share my projects with people who aren’t necessarily so technical. I also find that writing in a way that is understandable to the public really helps reinforce the concepts for myself.
Data journalism allows me to share my projects with people who aren’t necessarily so technical.
Every industry is become more reliant on data to tell their stories, and I think journalism is no exception to that. One thing that I love about the intersection of the two industries is that I think it allows for more honest reporting. Our understanding of events is often based on the perception we've already made, and while that bias still leaks into data science having hard numbers can usually help.
Shout!: So, in the end, can technology help predict suicidal risk, or the jury is still out?
Yoni Levine: Suicide is an extremely complicated problem that has many factors, and I don't think that this project or any others like it can fully account for all of these. I do however think that technology can help people in the healthcare domain scale to allow for quicker, smarter, and more prioritized care, which can definitely help the industry.
Shout!: Where can we follow your current and next work and can you tell us about it?
Yoni Levine: I’m about to start on an exciting new project to keep my skills sharp while I search for a job. The NIH recently made available over 100,000 anonymized chest x-rays, and I plan on using a machine learning technique called neural networks to automatically diagnose specific diseases in those scans.
You can find me on GitHub https://github.com/yonilevine/profile , read my blog on Medium https://email@example.com , reach out to me on LinkedIn https://www.linkedin.com/in/yoni-levine and sign up for occasional updates on my research here, tinyletter.com/YoniAtShoutNews.