Weekly Discussion Questions Based on Readings

Based on the readings each week, post a question for discussion in class. Try not to overlap with the questions of classmates.

Feb 6

• For instance: What might be some of the downsides to open-sourcing a piece of data journalism? - N. Diakopoulos

• Has the public effectively reciprocated the research/findings/work of data journalists? -Monique Robinson

• Nate Silver of 538 outlines their mission statement ("What the Fox Knows," March 17, 2014) and writes "The night our election forecasts are wrong will be a bad night for us." (footnote #1) Like all of the major polling outlets, 538 was wrong in their prediction about who would be President in 2016. What does this mean for a journalism outlet that prides itself its data mining and analysis when their predictions are incorrect? Will this mean a change for the mission of data-focused outlets like 538, or was the 2016 election an outlier in the data science? -Rebecca Gale

• In Nate Silver's article, he mentions how people like Brian Burke and Peggy Noonan have downplayed the validity of data in certain areas of sports and politics, respectively. Are these kinds of sentiments simply a result of not understanding what the statistics mean, or are there other more legitimate reasons for discrediting them? -Kyle Morel

• In the article of Computational journalism, it says "investigative reporters also expect plenty of false starts, tips that can’t stand up to scrutiny, and stories that rarely hew to the route expected at the outset. In short, they write stories, not studies." So, as a reader, how can I distinguish whether the news or reports are authentic or not since reporters can have false data in order to attract public's attention? -Xinyun Zhang

• Alberto Cairo in his article “Data journalism needs to up its own standards” points out that assuming certain incidents as a proxy variable of actual incidents is misleading as “Using news reports on kidnappings in Nigeria as a proxy variable of actual kidnappings is risky. You cannot assert that there are more kidnappings just because the media is running more stories about them. It might be that you’re seeing more stories simply because news publications are increasingly interested in this beat. “Can this type of misinformation be attributed to selective modern data journalism whereby it deceives its readers? –Sanchari Chowdhuri

• Philip Guo's article on data science workflow talks about data gathering, cleaning, analyzing and reporting. At times, the data you acquire might have outliers which are erroneously introduced or genuine instances. These outliers can potentially skew your reporting and result in inaccurate decision making. How should a data scientist (or even a data journalist) decide when such outliers are an important piece of information and when they should be discarded? - Sohan Shah

• Nate Silver in his article “What the Fox Knows” mentions that “the problem is not the failure to cite quantitative evidence. It’s doing so in a way that can be anecdotal and ad-hoc, rather than rigorous and empirical, and failing to ask the right questions of the data.” So as a data journalist when do you know that the question you have chosen for analysis is not the right question? The author also mentions generalization as a fundamental concern of science that is achieved by repeated experiments. How do we conduct repeated experiments to verify our hypotheses for a case like ‘Presidential Elections’ that happens once in 4 years? - Shashank Kava

• In "Clarifying Journalism's Quantitative Turn", Mark Coddington provides a typology that defines "data journalism", "computational journalism", and "computer assisted reporting". In real life news operations, how does this typology play out in terms of staffing? That is, in a given newsroom are there separate "data journalists" and "computational journalists"? How traditionally have journalists trained in "computer assisted reporting" branded themselves? -- Elliot Frank

• In Cairo's article about data journalism, he points out the problems associated with using proxy variables. His example is that you cannot assume that more kidnappings are happening in Nigeria simply because there are more news stories about kidnappings in Nigeria. Are there times when using proxy variables would be okay? What if you were graphing the change in news coverage of certain topics/beats?

• From the article by Alberto Cairo he states that data journalism may already be in a crisis because FiveThirtyEight and Vox underdelivered. We as journalists already have some mistrust with the public, so to be mistrusted by other journalists as well is an issue that will hinder data journalism. If we want to be trustworthy and do in depth analyses, how do we know how much research is enough? When do we know that we've used every source and gotten every fact about a topic? -Arielle Dupiton

Feb 13

• Celeste Lecompte's article about automation in the newsroom discusses the need to pre-assign blocks of text and check for sentence flow in order to adhere to AP style and ensure proper journalistic writing overall. With this in mind, how advanced can journalists realistically expect these data-driven stories to be in the near future? And is it worth it to focus so much attention on these algorithms when they still require human editing to ensure they are written properly? -Kyle Morel

• The article you co-authored with Tetyana Lokot observed news bot patterns and came to conclusions about these bots solely, from observation. These limitations are further discussed in the reading. However, why do you think the design decisions are important in understanding the news bots? To my understanding, what is observed can help you to gather all the information you may need about the methods of the news bot creators. -Monique Robinson

• The "dead squirrel" example Mark Zuckerberg speaks of in Celeste LeCompte's piece, "Automation in the Newsroom" highlights the divide between what editors deem important and what readers may find relevant. It's worth considering that more people are entering news sites through social media engines, such as a Facebook referral, rather than through the homepage of a website (i.e. www.nytimes.com). When print was the dominant form of media, editors could indicate importance through headlines, the size and location of the stories they wanted to prioritize. What does it mean for a news editor who wants to indicate importance when they no longer have a print medium, or even a homepage, to showcase what they believe to be the leading news? --Rebecca Gale

• In the guide to “Automated Journalism” by Andreas Graefe mentions that with the upcoming advancement in Natural Language processing and automated journalism, large number of news articles on similar lines are being generated. Since automated journalism is aimed at providing high customization and personalized news feed based on user location and previous user preferences. Does this pose a threat to users in terms of imparting biased and partial information and thereby influencing their opinions with a given set of factual data and representing only the partial picture through selective news feed? How can automated journalism address this issue of selective and opiated news reporting?-Sanchari Chowdhuri

• In the "Guide to Automated Journalism", Graefe describes two outlooks on algorithmic writing; man vs machine, and machine liberates man. While both views hold value, does the man vs machine model have potential to grow? That is, can we realistically foresee automated journalism taking the jobs of more journalists or has the system reached its peak in the newsroom? How much further can it go? - Amanda Smith

• In the article 'Domain Specific Newsbots', the authors describe how bots were used to generate news during the 2016 US Presidential elections and also the 2016 Olympics. I think the primary advantage of a bot over a human journalist is the ability to provide two-way real-time communication to the consumer. Are there other merits of bots that give them an edge over human journalists? -- Sohan Shah

• In the article 'Domain Specific Newsbots', the author discusses the need for a general purpose bot. By attempting to develop a bot for all platforms instead of making it domain specific, would the quality of the product be compromised? Also, if a general purpose bot is to be developed, the question then becomes which platform to develop for first? - Shashank Kava

• In your News Bots piece about automating news and information, there is always some skepticism when it comes to technology that tries to mimic the actions of humans. Yes a robot can do extensive research at lightning speeds but they lack some of the qualities that make us humane, for example emotions. How can we ensure that these bots will preserve the elements of newsworthiness, specifically impact? How or will they be able to connect with their audience to know what appeals to each person on a personal, emotional level? -Arielle Dupiton

In Automation in the Newsroom, it tells us how people are implementing automated news writing. But i still have some questions, like how can reporters and editors and programmers figure out every type of story to attract people. How can they adjust the bots to be not weird? What requires for them to figure out the purpose of the story and figure out a way of telling it? In my opinion, it must require very knowledgeable story telling skills. But, in this way, how do story-writing guys communicate with programming guys to successfully achieve what's in their mind about stories? --Xinyun Zhang

Feb 20

-In the article discussing 'social listening tools', the article mentions numerous tools that can sift through social media platforms to find potential news stories prior to their "fame." For example, the Dataminr scans tweets to access the tweets that could be the "next big story." However, in reading this article about this varying tools, I was interested in the statistics and findings of failed experiences with these tools wherein, a tool accessed information from a tweet or Facebook post that was false information. What degree do these algorithms take in sorting the fact from the fake? In some circumstances, it can be very difficult to distinguish the two. -Monique Robinson

I think that the change in editorial decisions to that of what garners interest (and thereore profit) versus what what warrants attention is a reflection of the sad state of much journalism today. News outlets are caught up in the chase of best analytics (how many clicks? views? redirects?), and as Columbia Journalism Review pointed out, this results in a decline in long-form, deeply analytic reporting in favor for "clickable" stories. - Helen Lyons

Another question on the Columbia Journalism Review's, "The New Importance of Social Listening Tools." According to the article, tools like Dataminr and News Whip have changed the role of editor. Instead of using editorial judgment to decide what people will pay attention to, the social listening tools render the editor as someone who will sort through what items "capture interest" to decide what is quality news. I wonder how this evolution of the news editor will be affected by newsrooms making additional cuts and preferring to invest in "clickable" stories over long-form (and time-intensive) journalism. Do news organizations feel an obligation to report on items that are more newsworthy (such as government) rather than click-worthy (such as kittens in a bathtub)? The social listening tools seem to prefer the latter. And furthermore, will editors be able to keep up with their new job descriptions? --Rebecca Gale

In Alexis Sobel Fitts' article, she discusses social listening tools and their ability to break news stories so quickly; however, in finding sources for the story, she mentions how some of the journalists she contacted requested anonymity due to "fearing retribution from their readers." What effect, if any, will readers' skepticism of this type of news-gathering technology have on its potential growth? In other words, if the average news consumer doubts the accuracy of social listening tools in obtaining information for stories, can it ever be more widely used and accepted in the future? --Kyle Morel

As mentioned in the article Finding the news lead in the data haystack: Automated local data journalism using crime data says “So in a way, social listening tools have simply shifted the role of the editor, from someone trying to figure out what will capture people’s attention to someone sorting through what we know will capture interest to finding what’s actually quality news. Those same editors also have to be diligent in rooting out misinformation and hyperbole.” Where in the news editors needs to exercise their prudence in order to determine articles and news worth reporting. However since quality of such news rely heavily on the raw data does this mean there are chances that these algorithms will only show news articles due to some trends going viral or false rumors rather than the actual news?-Sanchari Chowdhuri

Stray's article "The age of the cyborg" discusses how computational journalism can only go so far, and can not fully replace some of the work done by live journalists but instead helps make their job easier. What kind of problems can arise when journalists are also using computers to help create the story and run algorithms to find data? To what level should journalists be expected to fact check a computer system, and how responsible can they be if the algorithm contains an error? - Amanda Smith

The article "Finding the news lead in the data haystack" talks about automatically providing journalists with news leads from publicly available datasets. To improve the quality and accuracy of news leads provided, will machine learning be an effective solution? This will require the editor to provide feedback and report whether the news lead was sent to local newsrooms or not. Such feedback collected over time can be used as a training dataset for better predictive modelling. - Sohan Shah

How does the algorithmic process used by the tools described in "The new importance of ‘social listening’ tools" determine newsworthiness? How is this similar/different than traditional editorial determinations of newsworthiness? Also, what type of shortcomings could be consequences of the fact that the tool is biased towards working in English, even solely in the US (apart from the international example they use)- Elliot Frank

In Stray's article, it seems that "Izzy" is great for finding breaking news at alarming rates. I feel like the one advantage that a bot will have over humans is speed. But with technology comes technical issues and one of the main fields where error is can be detrimental is journalism. How can we assure that speed will not compromise the fact checking process so that all information is carefully analyzed for error? -Arielle

In the paper ‘Finding the news lead in the data haystack: Automated local data journalism using crime data’, the author mentions about automating analysis and monitoring large dataset at a local level to obtain potential stories and news leads. Is it worth spending all the resources on journalists as well as these automated tools and techniques to obtain such news leads for events that are relevant for a short span of time? -Shashank Kava

In the article "The Age of the Cyborg", it mentions that "Automated systems can report a figure, but they can’t yet say what it means" and "Reuters’s newest technology goes deeper, but with human help". In this way, what's the point of having those automated systems widely set up for journalism as it still needs people to review and change grammar mistakes and their narrative ways. I thought the main reason that people want to implement automated systems is that they want to save time and energy, producing more articles without much work. I can understand that people want to implement the system but why it needs to be broadly spread? --Xinyun Zhang

Feb 27

• The RevEx article mentions that the journalistic intent of this device and other devices similar to this is to help "capture user feedback and emotions more holistically, rather than just looking at stars and ratings”. However, do systematic programs and sifting tools like this truly grasp the humanistic aspect this tool suggests journalists are intending to get? -Monique Robinson

• The Propublica article mentions how Yelp users tend to give either 5 stars or 1 star in their reviews of restaurants and health care professionals. Is the service effective if most of the responses are so extreme? Is there a better way to measure user feedback? -Kyle Morel

• As mentioned in the article RevEx: Visual Investigative Journalism with A Million Healthcare Reviews “We noticed that frequencies can be misleading when the base rate changes dramatically between the categories of a facet. For instance, an item may have 90% of negative reviews on a total of 5 reviews only, whereas another item may have 70% of negative reviews on 1000 reviews. “Does this mean that effectivity of these software is limited by statistical limitation as it might be difficult to project a sentiment probability based on frequency? Will the misleading factor of frequency diminish with high sample ?- Sanchari Chowdhuri

--The ProPubicla piece has an interesting note: the negative Yelp reviews of health care providers complained of long wait times and a chilly bedside manner, rather than treatment or expertise. Yelp democratizes the reviewing process, allowing anyone to be a critic and to have a platform to broadcast their views. But what does this mean for fields like health care, where expertise and decision making are the crucial pieces, that reviewers focus on the experience? Interesting it's the feelings of dignity and respect that are the sticking points, and what does this mean for an office that focuses on the customer service aspect over expertise? --Rebecca Gale

The ProPublica article touches on the fact that even computers and data cannot account for human opinion or error. In this case of reviews, what are some ways an algorithm could help solve potential human errors in review so that businesses are not unjustly slammed, or is that not possible? - Amanda Smith

• The article on RevEx reads, "The best way to use these tools is to identify people and services that could be interviewed to gain more detailed information for a particular story". If it is up to the journalist to pursue a story based on the analysis results, could that choice of story be biased? A story backed with statistical results can also mislead an audience, especially when all the facts are not presented before them. - Sohan Shah

• The ProPublica article includes a chart which points out that for healthcare reviews on Yelp, 18% of the people gave a one-star rating and 65% of the people gave a five-star rating. Are these ratings by patients only based on factors like long wait times and difficulty of securing an appointment instead of the quality of treatment and diagnostic accuracy? Such generic rating system might be misleading for patients seeking medical attention. - Shashank Kava

• The article on ProPublica highlights the dangers of presenting information in a news story without a thorough explanation and critique of its value. What is the lay person's opinion of a doctor's performance worth? Can we really aggregate information like Yelp reviews, where the sources and their knowledge of what they are reviewing are unverifiable, and present it to an audience as being worth something? - Helen Lyons

• The journal paper was authored by NYU Engineering professors along with ProPublica journalists. The corresponding ProPublica article was co-produced with NPR's "Shots" Blog, and was born out of access to Yelp's reviews of healthcare providers. Does big data work necessitate/facilitate this type of collaboration, or has this always been a part of journalism? Is the RevEx tool designed so that a team can be working within the same corpus of documents?

--In "Visual Investigative Journalism with A Million Healthcare Reviews. Symposium on Computation + Journalism.", it says they use RevEx to explore and analyze millions of healthcare reviews obtained form Yelp. Is it really fair to rate healthcare data like restaurant rating data? I mean, those two data are so different. One is focused on human bodies while one is mostly for entertainment. Is that really practical to do that? --Xinyun Zhang

March 6

-In the “Facebook is eating the world” article, the writer suggests one of the best ways to combat ad-blockage is to force subscriptions and membership but what are some effective ways publishers can brand their platform so that users feel an affinity to purchase a subscription? -Monique Robinson

-Sifry's article talks about Facebook's role in possibly influencing voter turnout in elections by controlling the amount of political content it shows users. Should Facebook be analyzed in this way as a legitimate outlet for providing news? Or is it still primarily a social media site that should not be as focused and judged as much on political coverage it shows? -Kyle Morel

-Sifry's Mother Jones piece is such a piece of insightful reporting but it misses a big question: when (and not "if") will Facebook sell it's voter manipulation options to the highest bidder? The 2016 election proved that traditional campaign tactics aren't sufficient to win, so perhaps a digital algorithm, accompanied by an juicy ad buy, is the inevitable next step? (p.s. see the recent news of Snapchat and Everytown/Gun Safety) --Rebecca Gale

-As mentioned in the article “Bias in algorithmic filtering and personalization” and also in reality, not every Facebook friend is our actual friend and we segregate such acquaintances of ours in specific groups and share information accordingly. However in name of personalization, when the same Facebook data is being utilized by other applications they fail to understand the dynamics of the information and end up utilizing such information in an unwanted negative way like the Spotify case. What are the other negative effects of personalization and customization? Sanchari Chowdhuri

• I thought the "Building the Next New York Times Recommendation Engine" to be very interesting because it brings the audience into the decisions that are being made in the NYT Engineering department. The article uses graphs and explanation to break their process down, but for an audience who is not familiar with any machine learning or computer programming, my guess is that it would go over their heads. How responsible should a media/ social media company be to inform their readers/users of what editorial decisions are being made on an algorithmic level? An intuitive level? Obviously most companies will not publish explanations of proprietary algorithms, but even if they did, what percent of users would understand the implications? What percent would care?

• Sifry's article is even more relevant today after the 2016 election and the controversy surrounding Facebook's role in factually incorrect news and potential bias. The intersection between social media and journalism continues to grow and at the individual level is more easily monitored when it comes to fact-checks and biased reporting. However, can people really hold non-news corporations accountable to meet those same journalistic standards? If yes, how can that be done? - Amanda Smith

• Sifty's article points out that the monetization of information has given social media companies, like Facebook, a great deal of power over what content users see (or don't see). Given the enormous followings these companies have, is there a certain responsibility they have to ensure that information is accurate, varied, and useful? What does a private company owe to the good of the public? - Helen Lyons

• Micah L. Sifry's article highlights how Facebook conducts research experiments by manipulating users' newsfeed and also showing different views to different users. Is it important for a journalist to track such research when Facebook makes it public? Can the results of the research help a journalist or a news room with news dissemination? - Sohan Shah

• In the article "Building the Next New York Times Recommendation Engine", the author mentions a concept of using the simple average and the offset average for article recommendations. Can the concept be applied for recommending articles to individuals who read comparatively lesser number of articles and from varied domains? - Shashank Kava

-In the article about Facebook and voting, it reminds me of the movie Focus when they put an idea in someones head far in advance, without them knowing what they were seeing. Facebook did the same in planning years in advance to see who votes in order to determine the voter outcome of an election years later. How can this be used to gather other kinds of data, and how long do they need to collect data before having a good number of responses? Arielle

--In the article "Facebook Wants You to Vote on Tuesday. Here's How It Messed With Your Feed in 2012. ", it says that Facebook will promote voting by adding the profile button of "I'm voting" or "I'm a voter". I completely understand how it will promote people to vote by giving them social impact. However, can they manipulate people's voting choice as well? If so, by how? Will they just show what their friends vote? What if one day Facebook want to help one president candidate, will it really manipulate us? Won't that be too scary? --Xinyun Zhang

March 13

-I found the "Machine Bias: There's Software Used Across the Country to predict future criminals.And it's biased against blacks" article to be quite interesting. The most intriguing aspect was the high uncertainty of these varying machine produced risk scores and the decision of courts to integrate the results in their judgement, aside from the fact. Why would a judge choose to consider the output of an algorithm that has proven to be biased in cases? I know varying users have argued that they are not solely depending on this information but rather using the information as an informed guide; however, their decision-making is mind-boggling to me. I don't understand the rationale of essentially "listening to a voice" of a system that has generated ways to always put blacks at a disadvantage. -Monique Robinson

• One of the major benefits of online shopping is that it theoretically levels the playing field in shopping, and geography no longer becomes a factor. The information your computer stores on you can be used for informative purposes in journalism, and targeting purposes in advertising, but did it go too far when applied to sales? If so, what is the line for using that information to tailor the internet experience? - Amanda Smith

--I was so impressed with the Machine Bias readings and the ability to weave data sets into a compelling narrative. It was intriguing to me that this was published by ProPublica. Ostensibly, this was a weeks or months long research project, and one of the authors is listed as data scientist. Is this the future of journalism, where journalism is written collaboratively by organizations that do not rely on advertising revenue for funding? --Rebecca Gale

-ProPublica article: Putting aside the race issues with the Northpointe algorithm, is it logical to predict a likelihood of future criminal activity by strictly using computers, since criminal activity can vary from shoplifting to far more serious acts? -Kyle Morel

-As mentioned in the pro publica article “How We Analyzed the COMPAS Recidivism Algorithm” the COMPAS algorithm tends to be racially discriminating in its inference .Usually for a criminal to start redoing crimes a lot of parameters need to be taken into consideration like if the perpetrator was himself/herself a victim of such violence in past, to what degree did the assailant feel traumatized, socioeconomic factors etc. I am really interested to know how did the algorithm quantify factors like degree of impact of a violent incident on the perpetrator in past? If a person’s race was not mentioned did the algorithm give similar inference of the criminal’s probability to restart his crime activities? –Sanchari Chowdhuri

• “How We Analyzed the COMPAS Recidivism Algorithm” had me asking the same question I find myself asking often throughout this course and its readings: do the pitfalls of automated reporting outweight the ease of data analysis? Is the data analysis a crutch and a substitute for old-fashioned on-the-street reporting? To me, journalism is ultimately about people: human impact, human interest. The bias in crime data seems like a pretty clear example about how treating people like data points misses half the story.

I wonder if the change in data collection method will change the result of the study. I feel data should be collected by an agency and analyzed by other as well as completely removing the parameter of race as crime is independent of the race. Also, will the results change after working on a test data first and then the actual data? --Mayuresh Amdekar

• After reading the article "Websites Vary Prices, Deals Based on Users' Information", one question I kept wondering about is, how is offering different prices to different people legal? Also, the scenario mentioned in the article regarding differences in flight ticket prices is not the right analogy as the flight ticket prices for two individuals sitting in different geographic locations is never different. The flight ticket prices changes only with time and not any other factors. According to me the prices for a product should not vary based on the distance to a rival's store from the center of a ZIP Code and use of such personal information to modify parameters likes price should be illegal. –Shashank Kava

-- I'm reading the machine bias and I'm pretty surprised that the U.S. has the machine to rate for future criminals. I know it might be very useful to prevent bad issues from happening, but won't it be a little bit inhumane? Like, some people are judged by machines and they are even given some scores about their criminal potentials. Some people might commit crime due to a certain thing and reason and they might not do that again, will it be a little unfair to those people? --Xinyun Zhang

March 27

-After reading the "Peer Reviewing Our Data Stories" article, I questioned the articles mention of applying the peer-review method to sensitive stories. I was a little confused on the articles suggestion on dealing with this: " prepare a methodology piece that anticipates and addresses possible critiques but does not mention specific results." Related to this concern, I would like to know if peer-review has had a history of completely shaping the conversation of a data-story in sensitive topics? I think in discussing sensitive topics their is a high potential for opinion to be integrated and over-power the fact. - Monique Robinson

-Professor Diakopoulos and Koliska's article discusses the need for transparency in computational journalism used for news purposes. Because many readers are not familiar with this field of study, how do data journalists find a balance between providing readers with information and not including so much excessive material that it becomes confusing to the average person? -Kyle Morel

The Diakopoulos and Koliska article has a great soundbite: "transparency is the new objectivity," which sounds well and good but it's comparing apples to oranges, especially when it comes to political bias in reporting. A self-identified left-leaning or right-leaning publication will choose very different topics to cover, the bias begins with the initial question, "is this a story?" rather than "how do we cover it and what angle to take?" Transparency does little to change what is covered, and the initial bias is what will drive additional news coverage and conversations. Also, very impressed that the study collected 50 in-demand news media and academics for a day-long focus group without compensation...must have been a compelling invitation. --Rebecca Gale

As mentioned in the article Algorithmic Transparency in the News Media, results of Individual studies examining role of information disclosure and explanation of specific algorithms in recommendation systems, personalization and ranking enhances acceptance of recommendations and improve a user’s impression of recommendation quality, but also diminish the user experience, as Users are no longer attracted to the algorithm driving portals which understands their needs and recommends to them. Also many of these algorithms function so flawlessly because they capture data of user behavior in a state wherein users provide data about themselves without having to think how the same data is being used to manipulate recommendation to suit their tastes. However if the user is made aware how each of their data is effecting themselves with algorithm disclosure, Users might alter their data while providing thereby undermining the entire UI experience and recommendation systems. – Sanchari Chowdhuri

The article about peer reviewing data stories talks a lot about transparency in your data and methods, both before and after publication. Should that same standard be applied outside of data journalism? Is it realistic to expect journalists to share their process and notes for regular stories, and if not, why should they for data stories? - Amanda Smith

After reading the article "Peer Reviewing Our Data Stories" I believe, the peer review should be expanded to broader horizons by inducting the domain experts into it. Quite a few times, domain experts like in sociology, political science can offer better perspective into data generated as we as data scientists tend to get bogged down with numbers. I wonder what measures can be implemented to achieve this. -- Mayuresh Amdekar

The article 'Peer Reviewing Our Data Stories' talks about the open-sourcing of data analysis and also compares transparency in data journalism to academic research. I don't believe posting data and methods along with the story will be widely accepted by journalists. Competition is cut throat and it seems far-fetched to expect someone to share intellectual information, especially when their source of revenue depends on it. Data Scientists at Facebook are not obligated to share every finding of theirs publicly, why should there be a different set of "best-practices" or "ethics" drafted for journalists? - Sohan Shah

"Peer Reviewing our Data Stories" proposes that data journalism pieces should be held to a seemingly higher standard of review and editing. While I wholeheartedly support transparency in journalism, and would love to see more of it in the stories I read (both data and non-data stories), I wonder how common this is - does the average reader really want to know the nitty gritty back-end work that went into the story? Do too many details bore or confuse readers?

For "Peer Reviewing our Data Stories", I have a question for peer chosen for peer reviewing. Why data generators will always ask people with some extend of expertise to review data stories for them? I can understand that people know these techniques might have a wider overview about those stories and might be more eligible to edit and give some useful suggestions. But the most basic use of news story should be focused on the mass public, right? Why not ask normal people, just like some readers to review data stories, to see if they think they understand completely or if there's something ambiguous? ---Xinyun Zhang

April 3

-Monique Robinson: In the "Computational Journlaism and the Emergence of News Platforms" article, you discuss the strategies that will optimize the success of news industry. I think Buzzfeed does a great job of integrating these tools . However, I believe that more traditional media may have more challenges integrating each of these strategies because it will take away the value of news. How do you think traditional news outlets like NYT or Washington Post can apply the strategy described as a cultural shift of the "you" publish mindset, for example?

-Rebecca Gale: In the "Cultivating the Landscape of Innovation in Computational Journalism" article you discuss the effective brainstorming session in which students and media professionals came up with ideas. There seemed to be no shortage of innovative ways to deliver the news, but the discussion seemed geared toward a wide audience. How could news organizations effectively come up with innovative ways to deliver the news to their core audience, and how much more time can be spent for organizations to better understand who their core audience is? Rather than aiming for a "million clicks" how much more effective might it be to understand the target and effectively deliver?

-Kyle Morel: The article "Computational Journalism and the Emergence of News Platforms" references a Norwegian study that found computational journalism to often be more time-consuming and costly, contrary to what its proponents claim. With this in mind, what is the appropriate time for a relatively new media company to spend resources on this department? Should building a foundation in computational journalism be prioritized over simply building an audience first?

-Sanchari Chowdhuri : As mentioned in the article “Cultivating the Landscape of Innovation in Computational Journalism” the idea of using “3D virtual content for recreations of traffic intersections prone to accidents in order to help viewers get a better experience of why that spot could be dangerous and use of robotics for replaying sports highlights” appears to be very interesting and viable ideas to enhance creativity in journalism .There are high chances these ideas will instantly strike a chord of popularity among users as it just doesn’t report news but also allows users to deduce insights from the information based on his personal experience. This type of interactive news reporting is not possible in all fields. Will inclusion of such reporting techniques attract viewers to certain news genres which provide user interactivity while making other news genres less popular as they can’t offer such interactivity?

In the chapter "Computational Journalism and the Emergence of News Platforms", computational jouranlism is stated to be primed to create platforms and contribute reusable core technology assets to news organizations looking for competetive advantage. I wonder if this development can be taken up by journalism giants such as Washington Post, NY Times in collaboration instead of competetion. I strongly feel , that this exercise will definitely help in creating the desired platforms which then can be made open source to create scope of further development. Collaborative entry of big giants will give a constructive push to the development of field of CJ-- Mayuresh Amdekar

Amanda Smith: You mention the creation of new platforms in order to expand media's success - primarily Buzzfeed. How can traditional organizations that do not place as much emphasis on computational journalism contribute in a cost and time-effective manner to this movement?

In "Computational Journalism and the Emergence of News Platforms", it talked about the definition differences between CAR, DDJ, and CJ by reflecting their different coverage in story-telling aspect. Because of the computational thinking concept above, I'm wondering why we don't use this method to distinguish CAR, DDJ, and CJ? Basically, DDJ will be data thinking concept involved in the story while CJ should be computational thinking concept in stories. --Xinyun Zhang

April 10 -Monique Robinson: After reading Chapter 1: Newsgames, I questioned the appropriateness of applying games in news and conceptualizing the game, Cutthroat Capitalism, as a source for news that embodied all elements of good storytelling. I understand the use of word-puzzles and interactive quizzes by news organizations. I think these games can be actively integrated into differing story genres. However, how can a more active, game-like element be used in traditional news? The only way I can imagine a more interactive story is through visuals and graphics; however, it does not engage the reader into the experience like a video game and especially not like Cutthroat Capitalism.

--I enjoy reading Nate Silver and was glad to see chapters of his book as assigned reading. In chapter four, for years you've been telling us the rain is green, Silver gives us the real-world implications of failures in forecasting. The Schoolchildren blizzard in the late 1800s led to the creation of the National Weather Service, and though the failures of Katrina were many, but the forecasters held fast to their predications which proved to be correct. So what are other real-world implications of predictive analysis? Where are other predictive sciences(?) that have a public policy impact? --Rebecca Gale

-Sanchari Chowdhuri: As mentioned in news games chapter1, games like puzzles and quiz have been part of news for a long time, however I feel news worthy matter like attack from pirates is a serious issue and games on such topics might reduce the importance of such news. Although such games are a good way for readers to understand the ground reality of mode of operation of pirates and increase reader interaction; including such games in news medium is not prudent.

-Kyle Morel: In Chapter 4 of Nate Silver's book, he discusses the work of Edward Lorenz in the 1970s and how he discovered how inaccurate his weather predictions could be if there was just one small mistake. Here, the reason for the inaccuracy was not exactly a mistake, but simply a rounding choice by one of Lorenz's team members. Given that a variable like barometric pressure generally has an infinite number of decimal places that make it impossible not to round, is it realistic to think that weather predictions will ever be "perfect"?

-"The Parable of Google Flu: Traps in Big Data Analysis" - I don't think this should be looked at as a "trap" in big data analysis and seen as a failure of certain product features. Technology today is changing lives because people found large-scale errors 20 years ago and came up with innovative ways to solve them. The same applies to the fields of big data, machine learning and AI. Shouldn't we focus on learning what is being done to make these predictions more accurate rather than just what contributed to the errors? - Sohan Shah

-In the article "The parable of Google Flu: Traps in Big Data Analysis", it says "The core challenge is that most big data that have received popular attention are not the output of instruments designed to produce valid and reliable data amenable for scientific analysis." How does this happen? Why this would happen? The aim of big data analysis should be to spread more accurate and useful information, but why this situation will happen? I think we might need to look into this question so that it can improve the reliability of big data analysis. --Xinyun Zhang