Still remember the “read-one-book-in-eight-hours” activity?
Weeks ago, in Jackathon #3 , we turned our lab as a reading club , and had a fruitful day reading Data Science books. This time, the theme was to read something journalistic. 11 participants were required to finish reading one or two books in 8 hours, and share their findings or insights in the end.
Andy’s introduction: Which journalism?
Andy Shu introduced the books by illustrating an broadstroke outline of the intellectual disciplines that eventually lead to Journalism. He started with medieval universities, which usually house four “departments”: Arts, Medicine, Law, and Theology. The Arts department taught Trivium, i.e. Latin Grammar, Rhetorics, and Logic. It also taught Philosophy.
From “Traditional Philosophy”, that of Socrates and Aristotle, stemmed “Natural Philosophy”, the empirical study of the this world. Natural Philosophy gave birth to Science. Science developed and branched into three sub-fields: natural sciences, cognitive sciences, and social sciences. Natural sciences, such as Physics and Chemistry, study the physical world. Cognitive sciences and social sciences both study humans, the difference being cognitive sciences study the internal state of individual humans while social sciences focus on the interaction between humans.
Journalism roughly inherit from the Trivium tradition, paying keen attention to the power of words. Yet Journalism is also a trade. Journalism has a strong feeling of tradition and emphasize on the practitioners’ conscience and conduct. In this aspect, Journalism is similar to the craft of Law and Medicine.
Andy explained that the book list contains several scientific books because Journalism could learn a lot from the Science tradition - be it Cognitive Psychology, which studies human thinking and biases, or Probability Theory, which studies uncertainty. Click here to see the book list made by Andy and Tianyi.
Journalism Practice
Data Journalism Handbook
Yan Rong, a student majoring in mass communication, skimmed the Data Journalism Handbook and recommended it to beginners. The book introduced the workflow of doing data driven news by sharing many real-life cases in newsroom. Yan found that, for data journalists, the main task was to help audience understand data with cautious interpretation. Besides, she shared a case she found inspiring in the book. After the Fukushima nuclear disaster, Zeit Online, a German online media, simulated the situation when similar nuclear contamination happened in Germany, and made a map showing how many people would have to be evacuated then. “This project stood out, for it told a concrete data story and successfully engaged audience into a piece of news seemingly had no relation with them”, said Yan.
中國數據新聞工作坊培訓手冊
Kate Liu, UX assistant with Initium Lab, read the book 中國數據新聞工作坊培訓手冊. She found the book was quite helpful for freshmen. The book is a compilation of a training workshop in mainland China which taught the whole flow of data journalism, including searching and getting data from different portals, using Excel and Open Refine to clean and analyze data, and visualizing data with AI and code.
Verification handbook for investigative reporting
Verification handbook for investigative reporting tells readers how to get accurate information from sources of different credibilities. Tianyi picked this book and made a list of all tools mentioned in the book. She thought the skills applied to verifying user-generated contents are the highlight part. In a case, to verify whether a new war crime occurred in Nigeria, journalists examined a user-generated video frame by frame, extracting specifications of the road as well as details related to people, compare them with satellite images to check location. Additionally, they examined metadata in photos from different sources with EXif reader and did crosscheck. They even inverted picture color to see ID number printed on rifle in one frame.
Thinking Pattern
Signal and Noise
Sally Chong, a video journalist with rich experience, chose the book The Signal and Noise by Nate Silver, a statistician as well as founder of the data journalism blog FiveThirtyEight. The book tries to analyze how people predict future based on data about the past. She introduced three thinking patterns that would lead to flawed predictions.
The first error is out-of-sample, which happens when you make deduction from a sample that doesn’t contain relevant cases, hence, out-of-sample. One example is a drunk driver think he won’t have an accident, because he never had any accident driving sober.
The second mistake is overfitting, which occurs when your solution is only applied to the available data in specific, rather than all possible data in general; in other words, you try too hard to fit your theory to the current data.
The third issue to take note of is false-positive. This concept is better illustrated with an example. Let’s say 1000 people went for a cancer check, and some people are diagnosed as cancer carrier. Let’s assume the cancer check has a possibility of 90% to be correct. The question is, among these people being diagnosed with cancer, how many actually have cancer? Hint: it’s not 90% of them. Second hint: we don’t know yet.
To know the exact number of people with cancer, we need to know how many of the 1000 people are cancer carriers to begin with. Let’s say 50 of them do. So we start with 50 ill and 950 healthy.
The cancer check is 90% correct. That means 950 90% = 855 healthy people are correctly diagnosed as healthy; 950 10% = 95 healthy people are incorrectly labelled as ill.
Similarly, 50 90% = 45 ill people are correctly diagnosed, while 50 10% = 5 ill people are wrongly diagnosed as healthy.
So, in total, 95 + 45 = 140 people are diagnosed by the medical check to have cancer; yet, only 45 (32%) of them actually do, not 90%. The reason it’s not 90% is the cancer check marked a lot of healthy people as ill, due to the sheer number of healthy people. Those wrongly labelled people, called false-positive, “diluted” the final outcome. So, when we see a “positive” result, we need to think about how many of the positive results are actually positive and how many are false-positives. This effect is especially significant in rare events, such as having cancer.
Thinking, fast and slow
Thinking, fast and slow is another book about human thinking pattern. Nancy Hu found it filled with experiments and real-life examples. The book mainly targets to explore how rational people are when making decisions. The author found out that two totally different systems in human brain are driving the way we think. System 1 is based on intuition, emotions and previous experiences. It reacts fast and unconsciously, which is more like the default mode. Meanwhile, System 2 engages complex brain activity (like logical thinking) while working, but it only activates when System 1 gets stuck.
A Tradecraft Primer: Structured Analytic Techniques for Improving Intelligence Analysis
Pili Hu, data scientist with Initium Lab, chose A Tradecraft Primer: Structured Analytic Techniques for Improving Intelligence Analysis. “The book lists 12 practical techniques to improve structured analytics, with a core goal to train people think as rationally as machine”, Pili introduced. He made an analytic form for the 12 techniques, and found the “Analysis of Competing Hypotheses” the most hardcore one. The technique requires people to have more hypothesis and unbiased sample collection in decision making.
Writing
On Writing Well
First published in 1976, On Writing Well is a classic guide for writing news. Natalie Lung, majoring both in journalism and computer science, picked this book. The book can be divided into four parts: principles of good writing, methods to polish up, guide for writing different forms and mind-setting tips. The author offered a bunch of practical tips, for example, he held that science and technology news should not be a story simply about science but a story about people doing science. Besides reporters, the book is also good for people who have interests in writing or blogging. As the author encourages, “you write for yourself”, just write the things that attracts you and familiar to you, and anyone could become a good writer.
The Sense of Style
The Sense of Style is a relatively recent writing style guidebook which was published in 2014. “I think the book is responding to old school journalism textbooks like the Elements of Style (1920), and On Writing Well (1976) just shared by Natalie as well. The author pointed out a lot of new values which are contrary to tradition but well functioning in contemporary society”, said Guo Yanan, a social science major student.
Textbooks used to teach people strictly follow grammar rules, but The Sense of Style held a view that only a minority group of people still care that much. Nowadays young people are making and accepting new language conventions. In addition, the author encouraged writers to use a diverse and “unusual” vocabulary, especially the words would echo it in their sound and articulation. As for the “writing for yourself” theory, the book shares a quite different view with On Writing Well. “Don’t confuse an anecdote or a personal experience with the state of the world”, as the author put, suggesting trend stories are of more significance than stories about personal experiences.
History
News People Serving the Country
Victoria Jin, Initium Lab’s interactive designer read this one. This book consists of 15 papers. The papers give an account of Chinese news people’s awkward situation during the period of Republic of China. Some main discussions are highlighted in the book: the development of public sphere, the contradiction of liberalism, the standards of good news.
The public sphere is an area where individuals get together to freely discuss and identify social problems. Qing yi (清議) is regarded as China’s earliest public sphere. However, comparing to western world, public opinion in Chinese society lacked consensus, rationality and tolerance due to weak middle class and absent civic consciousness, and also because of the breaking down of family as an institution and its incapability as a buffer area between public sphere and individual.
The development of newspaper is built on the philosopher of liberalism, which also faces an embarrassing situation in China. For news people, national independence and modernization are two contradictory but equally significant objectives, that’s why liberalism get criticisms from both left and right parties.
A missing squire group also makes the scholars to play an important role in history. Some important names are mentioned in the book: Hu Shi realized that media should be impartial, She Chengwo promoted the civilization of media, Chen Leng insisted that newspaper should be professional.
Finally, there are strict standards to evaluate the quality of news at that time. Firstly, newspaper should be accurate with researches, allusions and stereotyped expressions. Secondly, a good story should be fulfilled with analysis and interpretation so as to connect fragmentary incidents in history as a whole. Finally, newspaper materials should be consistency in logic.
Reading Methods
How to read a book? (Chinese version)
How to Read a Book introduces a structural way of reading. According to this book, reading is divided into four levels - elementary reading, inspectional reading, analytical reading, and syntopical reading. The latter two are the focus of this book. In Frank Lin’s opinion, this book is a ‘must read’ for inexperienced readers since it builds your ‘reading sense’ from the ground up, while experience readers may find it too detailed.
What is Jackathon?
Jackathon is short for “Journalism-Hackathon”. At Initium Lab, we do data journalism, try digital storytelling and push limits of the news industry with technology. We organize regular Jackathons to advance our knowledge and skills in both journalism and technology.
Edited by Chao Tianyi