全部版块 我的主页
论坛 计量经济学与统计论坛 五区 计量经济学与统计软件 LATEX论坛
1930 2
2016-11-07

Learning and the future. Looking at top 10 most-watched videos containing the words ‘Data Science’ and/or ‘Data scientist’ in the title, published in the last 2 years, it stands out quite clearly that at the moment the main questions behind the most common queries revolve around: ‘how to become and what a data scientist is’ and ‘where data science is going’.

Screencastings and conferences still rule the style of communication whereas there is an evident lack of documentaries and more entertaining video contents that try to break-trough a non-technical, broader audience. The average length is around 65 minutes whereas the average number of views is 85,732. Even though the number of views may not be the only performance index, as a benchmark, in 2015 the average views per YouTube video of Science&Technology was 6,638, while Educational videos had 4,872 views.

1. Introduction to Data Science with R – Data Analysis Part 1 – (views: 264K)

Category: Tutorial

David Langer, Senior Director, BI and Analytics at Microsoft, holds a series of in-depth, hands-in tutorials on data science using R, going through the evergreen of Kaggle’s competitions: Titanic 101. The visuals, from a communication point of view, are simple and effective for the purposes of the video: a classical screencasting with audio narration.

2. Predictive Modelling Techniques | Data Science With R Tutorial – (views: 122K)

Category: Tutorial

The 3 hours, 10 minutes and 35 seconds tutorial offered by Simplilearn holds the second place of this most-watched top-10. It is quite impressive considering that the average length of the top 50 YouTube videos is roughly 3 minutes. The video is the fourth part of a series of tutorials that offer both a hands-in approach with some explanations on the theory behind predictive modelling. It covers the main types of regression models and some of their applications through some case studies. A slow-paced, screencasting with a narration (that sometimes might result a bit monotone), but perfect for beginners. The full ‘Data Science with R Language Certification Training’ course by Simplilearn is available here.

3. Data Science Tutorial for Beginners – 1 | What is Data Science? | Data Analytics Tools | Edureka – (views: 90K)

Category: Tutorial

The third place is hold by another tutorial on Data Science using R, Apache Mahout and Hadoop framework. This first part of the series of tutorials hold by Edureka!, gives mainly a more speculative introduction to Data Science (what Data Science is, the problems it tries to solve and prospects), Hadoop framework, R and machine learning using Mahout, ending with a more hands-in approach. 2 hours, 32 minutes and 55 seconds. Most watched videos on Data Science definitely seem not to follow the typical YouTube golden-rule for length.

4. “Data Science: Where are We Going?” – Dr. DJ Patil (Strata + Hadoop 2015) – (views: 83K)

Category: Panel

“Understanding and innovating with data has the potential to change the way we do anything for the better” – President Obama.

Dr. DJ Patil’s speech at Strata + Hadoop O’Reilly’s conference starts with the special message of President Obama advocating Data Science. This is the first video in the top-ten that delivers a broader picture on Data Science, its culture and possible impacts rather than a “how to” tutorial. The style is definitely engaging, with Dr. DJ Patil highlighting some critical points on how to unleash the full potential of Data Science in business and society. A must-watch if you are interested into “the bigger picture”.

5. Data Science – Part I – Building Predictive Analytics Capabilities – (views: 72K)

Category: Educational

A video lecture in a series about “how to build a modern analytics approaches” for individuals and professionals who don’t have a background in Data Science. The style is simple and clear: a 1 hour and 52 minutes of screencasting with audio narration. The video is an overview of the models and techniques related to Business Intelligence topics, predictive analytics and big data technologies (i.e. Hadoop) with some cases to show the potential of analytics in a business context. The slides are available here.

6. How To Become A Data Scientist — SF Data Science – (views: 65K)

Category: Panel

The video is the recording of a presentation held by Ryan Orban of Zipfian Academy and Dennis O’Brien of Idle Games “becoming a data scientist”. Ryan Orban offers an overview of what Data Science is, why the need of it, the different possible pathways to follow in order to become a Data Scientist (MS/PhD in Data Science, internships, self-study and immersive programs) and the disciplines embedded in Data Science. Dennis O’Brien’s presentation is about “What it’s like being a data scientist in a small startup”, sharing his learnings and insights with the audience.


二维码

扫码加我 拉你入群

请注明:姓名-公司-职位

以便审核进群资格,未注明则拒绝

全部回复
2016-11-7 12:25:21

7. Data Science – Eindhoven University of Technology – (views: 58K)

Category: Advertisement

A dramatic advertisement by Eindhoven University of Technology on their new programs in Data Science. The communication style is cinematographic, using computer graphics to mimic a sci-fi action movie, resembling to Minority Effect and Inception, through the use of technologies like Oculus rift and multi-touch interfaces. Here are the links to their Bachelor and Master programs.

8. Gigs: A day in the life of a data scientist – (views: 57K)

Category: Educational

RCRtv takes a look at a day in the life of a data scientist at the AT&T Foundry in Plano, Texas.

The video is an interview to the Senior Data Scientist Karthik Rajagopalan that answers to the questions: what data science does, background, how to stay on the cutting edge, tools adopted and some final advice for aspiring data scientists.

9. Data scientist vs Data analyst , their roles and qualification – (views: 47K)

Category: Educational

A screencasting video from Bigdata Simplified about the difference between data scientists and data analysts. The video stresses a lot on how the data is generated, its value and how to unleash its power, ending on what characterises a data analyst and what instead is peculiar of a data scientist.

10. The Future of Data Science – Data Science @ Stanford – (views: 37K)

Category: Educational

Dr. Euan Ashley, Associate Professor of Medicine & Genetics at Stanford University, Dr. Vijai Pande, Professor of Chemistry at Stanford University, Dr. Hector Garcia-Molina, Professor of Engineering & Electrical Engineering at Stanford University and Dr. John Hennessy, President of Stanford University, all together for an almost 26 minutes interesting conference on Data Science. The questions they try to address are, i.e.: how real is this emerging discipline? What opportunities and challenges does it present? How can Stanford nurture data science in research and education?

Observations:

Recently, I had the chance to talk to different data science startups. Their problem is: how reaching out their customers, alias business people, VC etc.

So, at this point I’ll translate my point of view in some questions:

  • Is it possible to talk about data science only referring to data science? The most watched top-ten videos are definitely not for a tech-illiterate audience, but for those who are already fertile to know more about this field. My idea is instead: why don’t we pivot the discussion on much more hot topics, such as policy making, urbanism, security, health, art and privacy?
  • Instead of focusing on “convincing” non data-driven companies to adopt a data-driven culture, hence to hire data scientists, why don’t we focus on translating data science to the masses first?

No matter how hard, on the internet, even at the bootcamp, they were stressing the importance of storytelling in Data Science, I felt that communication is a big issue in this field. And I’m not talking only about the visualization and reporting. It is old, it is clustered and often self-referential. I am aware I may sound bold and even a bit harsh, but what moves me to be this honest is a real desire to open Data Science up to the world. A key-performanceparameter should be how many tech-illiterate people get to know and familiarise with this field. Data Science might not be the holy grail, but it can change the world.

As a direct consequence of this thought, I decided to retrieve and analyse the text metadata of the first 500, most viewed videos on YouTube, from 2014, containing the words “data science” and or “data scientist” in their title. Once collected all the metadata, I concatenated in a bag of words all the tags, descriptions and title. I then used word2vec from gensim python framework. The analysis is simple but the results are intreresting.
The great majority of the 30725 words from the metadata are related to “business” (job, career, working, management, industry etc.), “education” (university, courses, learning, tutorial, programming etc.) and “data science tools and techniques” (machine learning, algorithms etc.). Words like “Future” and “Social” are only 0.13% of the total. The most similar words to “Future” are again related to “business” (job, career, information, product, chief). “Journalism”, “Progress”, “Political”, “Psychology” are less than 0.016% of the terms adopted when it comes to videos about Data Science. It is a field that also doesn’t seem to be “celebrity-oriented”. Dr. DJ Patil is the only influencer “to stand out”: 0.1% of the total number of words.

The picture that might come out from this very simple analysis is that:

  • Data Science is a clustered topic. Those who already know about it look for it. Especially for tutorials and educational purposes. The others don’t.
  • Data Science at the moment (from a communication point of view) strictly relates with careers/business.
  • There is an evident lack of reframing Data Science into “hot topics”.

One of the discussions I was having with a senior data scientist was about explaining the reliability of Data Science for business. I liked his expression “that Science should remind us of the Galilean experimental scientific method. In this field we experiment, a lot. It’s really hard to explain this point.”

We should also experiment with communication.

Bio: Marco Nasuto is a Data Scientist, Aerospace Engineer and Filmmaker. Now working in Denmark.


二维码

扫码加我 拉你入群

请注明:姓名-公司-职位

以便审核进群资格,未注明则拒绝

2016-11-7 12:52:48
oliyiyi 发表于 2016-11-7 12:24
Learning and the future. Looking at top 10 most-watched videos containing the words ‘Data Science’ ...
谢谢楼主分享的资料不错啊!
二维码

扫码加我 拉你入群

请注明:姓名-公司-职位

以便审核进群资格,未注明则拒绝

相关推荐
栏目导航
热门文章
推荐文章

说点什么

分享

扫码加好友,拉您进群
各岗位、行业、专业交流群