<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Eliecelnikier on Medium]]></title>
        <description><![CDATA[Stories by Eliecelnikier on Medium]]></description>
        <link>https://medium.com/@eliecelnikier?source=rss-2715e2ab6db8------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*cJUtuapbGSfWiUAtYWTrPQ.jpeg</url>
            <title>Stories by Eliecelnikier on Medium</title>
            <link>https://medium.com/@eliecelnikier?source=rss-2715e2ab6db8------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 18:03:51 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@eliecelnikier/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Data Science in Healthcare: what is it and why you should go for it.]]></title>
            <link>https://medium.com/@eliecelnikier/data-science-in-healthcare-what-is-it-and-why-you-should-go-for-it-6b3192357dc2?source=rss-2715e2ab6db8------2</link>
            <guid isPermaLink="false">https://medium.com/p/6b3192357dc2</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[education]]></category>
            <category><![CDATA[jobs]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[health]]></category>
            <dc:creator><![CDATA[Eliecelnikier]]></dc:creator>
            <pubDate>Thu, 15 Sep 2022 18:23:01 GMT</pubDate>
            <atom:updated>2022-09-15T18:23:01.676Z</atom:updated>
            <content:encoded><![CDATA[<h4>A review of data science in healthcare and why you might just find your new career path.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*24BefZmaOvmWrqM-" /><figcaption>Photo by <a href="https://unsplash.com/@thisisengineering?utm_source=medium&amp;utm_medium=referral">ThisisEngineering RAEng</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Today, the volume of <strong>digital data in healthcare represents 30% of the total volume of the data generated in the world</strong>, exceeding 2500 exabytes (knowing that one exabyte would store a 300,000-year call with your mom). This quantity has increased 15 times in the last eight years, making this sector the largest producer of data, well ahead of the media, industry, and finance sectors. By 2025, 453 exabytes of data will be produced and stored every day.</p><p>In healthcare, this astronomical amount of information is used to <strong>prevent the onset of diseases, diagnose more efficiently or find new treatments.</strong></p><p>But where does the health data come from? How can it be analyzed? What is the role of data science and why does the <strong>Harvard Business Review consider Data Scientist as the</strong> <strong>“sexiest job of the 21st century”</strong>?</p><h3>I. You are a data mine.</h3><p>The sources of data in healthcare are numerous. However, we can focus on three major domains: <strong>medical imaging, physiological data, and genetics.</strong></p><p><strong>Medical imaging </strong>(radiography, magnetic resonance imaging, scanner, microscope imaging…) accounts for 90% of health data. The analysis of this data allows the detection of breast cancer, tumors, bone problems, or the presence of malaria in cells with an accuracy of between 90 and 95%. This field uses a technique that we call <strong>computer vision. </strong>The principle of this technique is to give a large number of images to a computer so that it can learn to recognize pathologies by identifying common patterns between the images. Just as a child learns to distinguish a cat from a dog, the machine can also identify these species after being fed many examples of images.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*xLwTkk7Oq1OrW1C4.jpg" /></figure><p>The second-largest category of data is<strong> physiological data</strong>. According to MIT, <strong>our body produces 2 terabytes of data per day.</strong> This data, such as blood pressure, heartbeat, body temperature, glucose levels, and neural signals, is collected through sensors placed on our bodies <strong>The sale of these sensor devices has doubled in 4 years, from 73 million in 2016 to 161 million in 2020</strong> in a market currently dominated by Apple (100 million Apple Watch sales in 2020). These connected devices are transforming modern medicine by enabling continuous, precise, and personalized monitoring, guided by intelligent and automated analysis of the data collected.</p><p><strong>The diagnosis and tracking of patients with chronic diseases</strong> (diabetes, obesity, depression, Parkinson’s, etc.), the <strong>continuous monitoring of biomarkers</strong> (weight control, mental health assessment, addiction monitoring, etc.) and the <strong>prediction and prevention of diseases</strong> are all aspects of medicine that will be transformed by large-scale data analysis. Several companies in the United States, such as <a href="https://goforward.com/?g_network=g&amp;g_adid=504375283346&amp;g_keyword=forward%20health&amp;g_adtype=search&amp;g_adgroupid=68969552639&amp;g_keywordid=kwd-375370153160&amp;g_campaign=G-Brand_NY0&amp;g_acctid=488-286-6060&amp;g_campaignid=1848532062&amp;utm_medium=cpc&amp;utm_source=google&amp;utm_campaign=G-Brand_NY0&amp;utm_content=forward%20health&amp;utm_term=forward%20health|e&amp;gclid=Cj0KCQjwhqaVBhCxARIsAHK1tiNHXKmRkfAfxCiOIpu4CiyWvuSJywhHM5F3hAyWucNFlLuNY1rFEtYaAtsUEALw_wcB">Forward Health</a>, have already created several medical centers where all measurements are automated and processed with intelligent algorithms, from taking temperatures to detecting skin cancer to assessing mental health. In Uganda, a country where infant mortality exceeds 7%, <a href="https://neopenda.com/">Neopenda</a> develops connected bracelets, that measure vital signals of newborns, in clinics that lack qualified personnel. These clinics can then rely on this technology to monitor hundreds of newborns simultaneously.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*fLVsQh_SfHHNSReI" /><figcaption>Photo by <a href="https://unsplash.com/es/@dawson2406?utm_source=medium&amp;utm_medium=referral">Stephen Dawson</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Finally, with more than <strong>10 billion kilometers of DNA and 10 trillion cells in each human body</strong>, we represent an exceptional pool of valuable <strong>biological information.</strong> Predicting the mutations of our cells and the structure of our proteins solely through our genetic code enhances biological knowledge and dramatically accelerates the discovery of new, more precise, and more effective treatments. The company <a href="https://www.deepmind.com/">Deep Mind</a>, acquired in 2014 by Google, has just revolutionized this field by finding the solution to a 50-year-old problem: <strong>predicting the 3D structure of a protein from an amino acid sequence (ACGT)</strong>. In 1969, American biologist Cyrus Levinthal estimated that using brute force by enumerating all possible structures of a protein before finding the right one <strong>would take longer than the current age of the universe. Today, </strong><a href="https://www.deepmind.com/blog/alphafold-using-ai-for-scientific-discovery-2020"><strong>AlphaFold</strong></a><strong>, the model involved here, takes only minutes.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TF16-NHUeyh-7ULUeFFGgA.png" /></figure><h3>II. The AI to prevent, detect and cure</h3><p>As we have seen, the <strong>amount of data is staggering and its expansion is exponential</strong>. In response to this abundance, a new science was born in the early 90s, and without any surprise, it is the data science. A science with many aspects and many definitions.</p><p>To understand the context in which this science was born, let’s start by defining some terms. <strong>Artificial intelligence has existed since the 1950s and consists of imitating forms of human intelligence.</strong> But contrary to popular belief, computers are still far from reaching the complexity and power of the human brain (the most complex tool in the universe to date). <strong>There is still a factor of 10⁶ in terms of computing power between the most powerful processors and the human brain, capable of 10¹⁶ operations/second.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/796/0*NPhOGWr42IQ9mkX0" /><figcaption>Power vs memory for different systems</figcaption></figure><p>On May 11, 1997, the <strong>AI DeepBlue defeated Gary Kasparov, the world chess champion, in 37 moves.</strong> The power of computers has increased 520 times since 1997 and continues to double every 18 months on average. Predicting a large number of moves is a priori a matter of human intelligence, of course, but it is the computing power that has made this possible. <strong>But then, what are the fields of AI that have really progressed these recent years? Are there areas that are closer to human intelligence and that do not rely solely on computing power?</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*YZlcSMa9dYGzT9XU" /><figcaption>Photo by <a href="https://unsplash.com/@jeshoots?utm_source=medium&amp;utm_medium=referral">JESHOOTS.COM</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p><strong>Machine learning, a sub-domain of AI, is the domain par excellence that uses this computing power without relying solely on brute force. </strong>Machine learning reproduces human behaviors in an autonomous way. It is said that the algorithm learns to decide by itself by modeling a particular situation after having been confronted with a large number of similar experiences.</p><p>Machine learning can be used to <strong>detect which antidepressant will work best in a patient using neural signals</strong> (<a href="https://www.altoneuroscience.com/">Alto Neuroscience</a>), <strong>detect whether someone is cycling or running through inertial sensors</strong> (Apple Watch), or <strong>predict which movie you are likely to like</strong> (Netflix). Deep learning, the second layer of Machine Learning, uses an additional degree of abstraction in learning decision models. It is said to be deep learning, hence the name. Deep learning models can <strong>predict the severity of Covid</strong> in a person based solely on a chest x-ray (<a href="https://owkin.com/">Owkin</a>), <strong>summarize a multi-page article in a few sentences</strong> (<a href="https://huggingface.co/">HuggingFace</a>), or <strong>track and understand the gait patterns of patients with neurodegenerative diseases </strong>(<a href="https://www.evolutiondevices.com/">Evolution Devices</a>).</p><h3>III. A job at the intersection of several fields</h3><p>Data science could be defined as the <strong>combination of several disciplines such as statistical analysis, predictive analysis, visualization, mathematics, computer science, and machine learning</strong>. The data scientist is the one who has to play with these fields with one objective in mind: to process the data in order to bring value to it. He is the contemporary oil refiner, and data is the petroleum of the 21st century.</p><p>One should not be afraid of the number of disciplines that this profession includes, on the contrary, it is one of the rare professions where one learns constantly. <strong>It is possible to learn almost everything online.</strong> More and more high-quality courses are published. To begin with, the <a href="https://fr.coursera.org/learn/machine-learning">Stanford course on Coursera by Andrew Ng</a> is the most educational and complete free course on machine learning. After this course, you will be able to understand exactly what machine learning is, to code your first intelligent algorithms <strong>capable of recognizing a cat in a photo or predicting the next movements of the S&amp;P 500.</strong></p><p>It is also a job with different definitions. Some data scientists at Google have a role close to that of a manager, others can classified as software developers, others deal exclusively with machine learning, and still, some only perform data visualization. <strong>We talk about data science as a profession in itself, but it is first and foremost a palette of very powerful tools that you have to combine and use in the field that you’re passionate about.</strong></p><p>My choice was health and neuroscience and I would be happy to talk about it in a future article. If you are more interested in industry, ecology, fashion, architecture, or sports, these are all fields that will be heavily transformed by data science as well.</p><h3><strong>Don’t put barriers in your way, choose your field, and go for it.</strong></h3><h3>Reading list:</h3><p><strong>Articles</strong></p><ul><li><a href="https://deepmind.com/blog/article/AlphaFold-Using-AI-for-scientific-discovery">Alpha Fold : Using AI for scientific discovery</a></li><li><a href="https://arxiv.org/abs/2007.03606">Data Science : A comprehensive Overview</a></li><li><a href="https://hbr.org/2012/10/data-scientist-the-sexiest-job-of-the-21st-century">Data Scientist: The Sexiest Job of the 21st Century</a></li></ul><p><strong>Books</strong></p><ul><li>Bad Blood, John Carreyrou</li><li>Apocalypse Cognitive, Gérald Bronner</li></ul><p><strong>Online Courses</strong></p><ul><li>Learn the basics of ML : <a href="https://fr.coursera.org/learn/machine-learning">Machine Learning, Stanford, Andrew Ng</a></li><li>Dive in Deep Learning : <a href="https://fr.coursera.org/specializations/deep-learning">Deep Learning, Andrew Ng</a></li><li>More general AI : <a href="https://engineering.fb.com/2016/05/10/ai-research/ai-revealed/">Artificial Intelligence, revealed, Facebook Engineering, Yann LeCun</a></li><li>Go deeper into math : <a href="https://www.youtube.com/c/3blue1brown">https://www.youtube.com/c/3blue1brown</a></li></ul><h3><strong>References</strong></h3><ul><li><a href="https://www.statista.com/statistics/1037970/global-healthcare-data-volume/">https://www.statista.com/statistics/1037970/global-healthcare-data-volume/</a></li><li><a href="https://hbr.org/2012/10/data-scientist-the-sexiest-job-of-the-21st-century">https://hbr.org/2012/10/data-scientist-the-sexiest-job-of-the-21st-century</a></li><li><a href="https://www.nature.com/articles/s41746-021-00438-z.pdf">https://www.nature.com/articles/s41746-021-00438-z.pdf</a></li><li><a href="https://mitsloan.mit.edu/ideas-made-to-matter/patients-key-to-making-sense-medical-data">https://mitsloan.mit.edu/ideas-made-to-matter/patients-key-to-making-sense-medical-data</a></li><li>Lecun (2019)</li><li>Apocalypse Cognitive, Gérald Bronner (2021)</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6b3192357dc2" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>