1646-5954 1646-5954 S1646-59542018000300004 Spain 00 09 2018 00 09 2018 12 3 66 82

Contributions of academic articles to the practice of journalism and data management

 

Idoia Portilla*

* Universidad de Navarra, Spain

 

ABSTRACT

The increasing number of open data resources and the arrival of big data have boosted the data available as a source of news. Journalists need new skills for collecting the data and creating the news. In addition, all this data can also be used to provide new media services and to take media business decisions, and journalists need new skills related to data for these tasks. Taking into account these areas of knowledge required by journalists for the use of data, we perform a structured literature review (SLR) followed by a content analysis. The results confirm the relevance of data management in journalistic practice, requiring skills in statistics, data visualization, technology, but also in ethics, marketing or audience monitoring.

Keywords: skills, big data, management, content analysis, visualization.

 

Introduction

]]> The relevance of acquiring skills for data management in journalism practice has been boosted by the increasing number of open data resources and the arrival of big data (Elías-Pérez, 2015; Parasie, 2015). At the same time, the media capable of working with data can generate databases that can be sold or used to offer extra services (Gray, Bounegru, and Chambers, 2012). This is the case, for example, of ProPublica (www.propublica.org/datastore/) or Bloomberg for enterprise (www.bloomberg.com/enterprise/).

But data is not only a product or service for the media. Nowadays, all digital business generates a great amount of data, and the proper use of this big data can improve company results (Mayer-Scho¨nberger and Cukier, 2013; Lewis and Westlund, 2015). It is the era of big data: high-volume of data in high-variety information assets, that is generated at high-velocity and requires being analysed also very quickly to gain valuable insights. As other media companies with big data, media companies should learn how to monetize the data generated by the digitalization of their content, taking advantage of information such as the visits to their webs, the fans of their brand or the searches made in their cross-platform content services.

In response to these realities, journalists should gain new competencies related to the management of data. In this paper, we review the milestones linking data with journalism practice. The first step is obtaining the main areas of knowledge required by journalists in order to use data. The second step is performing a quantitative literature review of research contributions published in the main academic journals. The results of this review, complemented by a content analysis, will provide a list of areas of knowledge and skills that are becoming important requirements for the journalist in this era of big data.

 

Milestones related to data and journalism

Journalists have been using data in economic or sports news since the origins of media. The first periodical publication in the eighteenth century, The Lloyd's List, was a weekly journal of general commercial news and included price lists and details of ships arriving at English and Irish ports (Arrese, 2002). According to The Guardian, their first data journalism content was published in 1821 with a list of schools in Manchester and Salford, how many pupils attended each one and their average annual spending (Datablog, 2011).

Data continued being source and content of news in the following decades, but the arrival of computers provided a great step forward. In 1952, the US television network CBS used a computer to predict election results (Hewett, 2013; Uskali and Kuutti, 2015). In 1967, the coverage of the 1967 Detroit race riots by Phillip Meyer won the Pulitzer Prize (Meyer, 2001; Uskali and Kuutti, 2015). This last journalist is considered the first American journalist to have used computers to conduct investigative journalism (Parasie and Dagiral, 2013), and many authors reference his work in the history of data journalism (Hughes, 2013; Bradshaw, 2014; Elías-Pérez, 2015).

However, many journalists were not capable of analysing and reporting data, so they needed training. Meyer himself was the author of the handbook Precision journalism published in the 70's (Meyer, 2001; Coddington, 2015). In his book, he tried to show media professionals how to deal with data and surveys in their reports (Meyer, 2001). In the following years, we can find other handbooks about computer reporting (Houston, 1996; DeFleur, 1997; Garrison, 1998; Houston, Bruzzese, and Weinberg, 2002), showing the relevance of the issue. The set-up of the National Institute for Computer-Assisted Reporting in the United States in 1989 (Hughes, 2013; Parasie and Dagiral, 2013) and the Centre for Investigative Journalism (CIT) in 2003 in the UK (Hewett, 2013) also helped train journalists in data management.

From the mid-1990s, the number of data-oriented projects increased thanks to the growing availability of computerized data and the world wide web (Parasie, 2015). In 2010, The Guardian in the UK and the New York Times and the Chicago Tribune in the United States developed their own data journalism teams and started publishing news with the data they elaborated (Uskali and Kuutti, 2015). In 2012 the Global Editors Network presented the first Data Journalism awards (Uskali and Kuutti, 2015).

The educational resources teaching how to use data as a source, to collect it, to analyse it and to report it, have been included under several terms: precision journalism, Computer Assisted Reporting (CAR) or investigative journalism (Meyer, 2001; Coddington, 2015; Uskali and Kuutti, 2015). The term “data journalism” has generally been used since the mid-2000s (Royal and Blasingame, 2015). However, it is still difficult to define (Ferreras-Rodríguez, 2012; Beleaga, 2013; Coddington, 2015; Royal and Blasingame, 2015). Professionals and scholars sometimes separate data journalism from other practices such as computational journalism, programmer journalism or open-source journalism (Coddington, 2015).

]]> If we consider any kind of journalism that uses data as a source for its news and that publishes stories using data, we can divide its practice into three steps: access to the data, its analysis and the publication of the story (Ferreras-Rodríguez, 2012; Teruel and Blanco, 2015). If we consider the analysis as a skill necessary for reporting, two relevant areas of knowledge stand out over the rest: data as a source of news and the use of data in the content of news.

 

Open data and software

Other milestones related to data and journalism are the Open Systems Movement in the late 1970s (Holtgrewe and Werle, 2001), the open source principle born in the late 1990s (Coddington, 2015) and the open data policies (Hughes, 2013). These open movements have facilitated the access to data and the software for its analysis (Ferreras-Rodríguez, 2012).

The passing of the Freedom of Information Act in the United States was in 1966 (Bradshaw, 2014), and it gave people the right to access information from the government. However, the release of databases by government authorities, the actual open data policies, started in the 1980s (Parasie and Dagiral, 2013). In the UK, the Freedom of Information Act received royal assent in 2000; it came fully into force in 2005 (Davies, 2010; Hewett, 2013; Bradshaw, 2014) and, in 2010, the government open data website data.gov.uk was launched (Davies, 2010; Hewett, 2013), a few years after The Guardian journalists´ campaign called Free Our Data (Davies, 2010; Hewett, 2013; Bradshaw, 2014). In the European Union, regulation 1049/2001 is the transparency law, although other countries have already tried to regulate it: Sweden, in 1776, Finland in 1951, and Spain in 1992 (La-Rosa and Sandoval-Martín, 2016).

The Open Government Partnership was launched in 2011 (Bradshaw, 2014). At that time, governments began to become members so as to demonstrate their compromise with transparency and the empowerment of citizens (Open Government Partnership, 2016). Regulators, consumer groups, charities, scientific institutions and businesses also began releasing newsworthy data (Bradshaw, 2010), making the analysis of data among journalists more common.

But the use of government data for developing news content has been seen as the beginning of a new kind of data journalism, mainly due to Wikileaks. This release of war logs from Iran and Afghanistan in 2010 is considered a significant event for data journalism (Hewett, 2016), and even the starting point for contemporary data journalism activities (Uskali and Kuutti, 2015). In 2013, when Edward Snowden revealed that intelligence agencies were engaging in secret global surveillance, the debate about privacy moved to the foreground (McStay, 2017; Lewis and Westlund, 2015). Thus, the 2010s are relevant dates for data and journalism practice.

In brief, these open data and open source movements helped journalists in the accessing of data and its analysis for reporting a story, again bringing these two areas to the forefront, now together with the privacy debate.

 

Portals and services based in data

Besides data as a source, the open data movement provided portals and services based on data. In 1998 the UK’s mySociety developed Up My Street, a website with information of local areas (Bradshaw, 2014). In 2004, Washington DC launched its data portal (Parasie and Dagiral, 2013). These portals facilitated access to data for news, but also the development of new media services mainly as public services for their audiences.

Since 2005, journalists have been hired to design web applications, with the purpose of strengthening government accountability and citizen participation (Parasie and Dagiral, 2013). In 2007, the Knight Foundation awarded the project EveryBlock, a website with information about neighbourhoods. The project PolitiFact, a fact-checking website that rates the accuracy of claims by US politics, received a Pulitzer award for its journalist content based on data and analytics (Ferreras-Rodríguez, 2012).

Some data services, such as the BBC budget and fuel calculator, have helped UK citizens to control their expenses (Hurrel and Walton, 2013). Others, such as mashups and web applications, make data transparent to the audience and can be used as decision-making tools, for example, to decide where to live, knowing whether the neighbourhood is a secure area (Parasie and Dagiral, 2013). So, online data services are seen as an entrepreneur model of data journalism activities in the newsroom, according to Uskali and Kuutti (2015), with stories that incorporate user participation (Royal and Blasingame, 2015), and look to help people to become more powerful.

As we have seen in the introduction, some media, such as ProPublica or Bloomberg generate databases to be sold (Gray, Bounegru, and Chambers, 2012). There are also news organizations that employ data analytics as their niche, such as Nate Silver`s FiveThrityEight (http://fivethirtyeight.com/) in the United States (Royal and Blasingame, 2015), or Fundación Civio (www.civio.es) and Porcentual (www.porcentual.es) in Spain.

]]> Nowadays, the increasing number and use of mobile phones has forced media to offer their services via apps. News applications can personalize data to the user offering information about traffic or sports (Westlund, 2013; Howard, 2014; Silva, López, Westlund and Ulloa, 2016).

According to all these examples, apart from data as a source of news and as content of them, a third area to be considered is the use of data to offer media services, which can provide new revenue resources for media companies so journalists should be aware of it.

 

The era of Big Data

The development of big data is the last milestone we highlight. This term was coined in the 2000s by sciences such as astronomy and genomic, and it migrated to all areas about ten years later (Mayer-Scho¨nberger and Cukier, 2013).

Big data can be defined as the data generated through the use of digital technologies (Margetts, 2014) and since 2001 it has been described by three Vs: volume, velocity, and variety (Laney, 2001). It can also refer to the processes surrounding this great volume of data, which is too large for standard computer memory and software (Lewis and Westlund, 2015; Parasie, 2015).

Due to the increase in digital information which led to the big data era, data-driven journalism has become more prominent (Coddington, 2015). According to Elías-Pérez (2015), the computer experts know how to manage data, but the journalists are the ones who know what questions should be asked to tell the story. Therefore, big data affects data journalism in sources used and in reporting. Now it is easier to find the data, but the journalist needs to select, analyse and compare it and finally publish the news (Ferreras-Rodríguez, 2012).

Furthermore, big data relates to business. Digitalization has given media companies computer data that includes information about the user (Manovich, 2001; Napoli, 2016) and can drive the online business of the media (Aguado and Martínez, 2009; Lewis and Westlund, 2015; Royal and Blasingame, 2015). The most used methods of capturing data are online tracking cookies and website pathing (Forbes Insights, 2014) but also mobile apps can provide useful information (Howard, 2014). Data can be used to customize automatically the media composition, personalizing the offer (Evens and Van Damme, 2016). It offers great opportunities to the communication market (Marks, 2013; Papí-Gálvez, 2014). The media can implement big data strategies for a better understanding and targeting of customers (Napoli, 2016).

In brief, the big data era means that the management of data is essential for journalists because the data available is increasing and news stories should take this into account. However, media companies have also gained big data related to their business, so journalists should learn how to manage it for decision-making.

 

From the milestones to areas linking data and journalism

Taking into account the development of data journalism and the aforementioned milestones related to data and journalism practice, we will consider four areas that linked journalism and data:

  1. Data as a source of news.
    2. ]]>

    Data has been the source of news since the origin of the media but journalists received real training in the collection of data with the Meyer’s precision journalism book and the development of Computer Assisted Reporting (CAR). With the open data movements, access to data became easier, and nowadays, in the era of big data, journalists can find more and more data that sometimes require computational knowledge in order to retrieve them.

    1. Data in the content of news (reporting).

    Once journalists have the data, they need to analyse it and write the story. The educational resources under the names of precision journalism, investigative journalism, computer-assisted reporting or data journalism have tried to teach how to do this. In addition, the open software movement has helped by facilitating access to analytics and visualization software. In the era of big data, journalists have data to complement many of their news stories. 

    1. Data for providing new media services.

    The open data movement gave access to government data and media have used it to develop sites, portals and apps based on this information. With these innovative services, media have tried to facilitate access to data for citizens and obtain new revenues. The big data available can help to increase the number of these services.

    1. Data in the media business.
    ]]> Digitalization has provided big data for online media. Journalists and managers of the media have to use data to improve their products - by providing personalized news, for example. Like other companies with digital data, the media should try implementing big data strategies to increase their revenues.

    Once we have identified the areas that link data with journalism and journalists, the next step is to analyse whether these issues and milestones appear in academic publications, when and related to which skills or other issues. As open data and big data have boosted the volume of data available, the period considered in our research will include these two milestones.

     

    Method

    The methodology applied in this work is a structured literature review (SLR) followed by a content analysis with first, a priori coding and, in a second step, an emergent coding. The result is the combination of methods used in health and the social sciences.

    According to Dawidowicz (2010, 5), “a literature review is an examination of scholarly information and research-based information on a specific topic”. When it uses rigid rules is called a structured literature review (SLR) (Massaro, Dumay, and Guthrie, 2016). The use of rules offers less bias, more transparency, and the possibility of replication of the review (Massaro, Dumay, and Guthrie, 2016; Tranfield, Denyer, and Smart, 2003).

    The SLR is frequently applied to develop insights and future research paths or review methodologies applied (Moher et al., 2009; López-Cantos, 2015; Massaro, Dumay, and Guthrie, 2016; Ausserhofer et al., 2017). However, in this case the objective is translating areas of research into skills related to journalism practice.

    The process for the selection of contributions will be drawn using the flow diagram proposed by the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analysis) from the field of health (Liberati et al., 2009; Moher et al., 2009). This diagram helps in explaining the process of the identification of records, the screening, the eligibility and the final records included (Moher et al., 2009).

    After the selection of cases, the data will be coded to develop insights (Massaro, Dumay, and Guthrie, 2016). This codification follows the process of a content analysis (Wimmer and Dominick, 2011), registering items for each article (which constitute the unit of analysis). The content categories will be established using a priori coding for those that can be listed before the analysis. In a second step, the categories will come from an emergent coding, establishing categories after the examination of the articles (Wimmer and Dominick, 2011).

     

    ]]> Research questions

    The final purpose of this article is to know what can be learnt about skills related to the use of data in journalism from the contributions of researches published in academic journals. There have been previous literature reviews for identifying influential publications and gaps in the research of data journalism (Ausserhofer et al., 2017), for identifying methodologies (López-Cantos, 2015) or opportunities for innovation in computational journalism (Diakopoulos, 2012). However, this paper differs from these previous contributions in the method used for the selection of the contributions and the research questions made. Firstly, it extends the interest to any article that relates data and journalism/journalists. Secondly, the questions refer to the issues discussed by academia with the focus on the skills and knowledge needed in the practice of journalism.

    The research questions are:

     

    Universe of study and final sample

    The unit of analysis will be the academic contributions connected with data and journalism or journalist published from January 2006 to December 2016. This period includes the years of the development of open data and big data, the two milestones of interest. The key criterion for the selection of academic contributions is that they should include the terms “data” and “journalism” or “journalist” in their abstracts.

    ]]> The selection is made using the tool Unika, from the library of the Universidad de Navarra (www.unav.edu/en/web/biblioteca). This search tool has been in operation since September 2011 (www.unav.edu/web/vida-universitaria/detalle-noticia-pestana/2011/09/28/unika:-nuevo-buscador-unificado-de-la-biblioteca-de-la-universidad-de-navarra?articleId=302098). It provides access to several academic journal databases. In 2017 this tool included more than 350 providers such as Academic OneFile, BASE, Biography in Context, Business Insights (Essentials), Business Source Ultimate, Expanded Academic ASAP, General OneFile, General Reference Center Gold, InfoTrac Computer Database, InfoTrac Newsstand, InfoTrac Student Edition, NewsBank and Scopus.

    The Unika search tool provides information on the contribution including the title of the article, the authors, the journal, volume, issue and pages, the keywords (from the authors or by the database) and, in many cases, the abstract. The search took place on April 25 2017. The date is relevant because the journals included in each database can vary, and also the contract between the library and the providers can change over time.

    Unika permits an advanced search and the criteria considered were:

    In a first search, the fourth criterion was that the abstract should include the terms “data” and “journalism”. As a result, we found 696 articles. In a second search, the articles should include the terms “data” and “journalist” in the abstract, replicating the other criteria, and we found 553 articles. The final number was 1,249 articles (see Figure 1).

     

     

    ]]> Due to the high number of articles and the fact that only one coder would analyse the information to avoid reliability limitations (Wimmer and Dominick, 2011), only the two top databases were selected (see Table 1): Communication & Mass Media Complete and Academic OneFile.

     

     

    The databases selected included 521 articles of the 1,249, 41.7% of the total (see Table 1). The Communication & Mass Media Complete website (EBSCO, 2017) includes full text for more than 440 journals, abstracts for more than 670 journals and selected coverage of nearly 180 journals. Academic OneFile provides access to articles from over 17,000 scholarly journals and other authoritative sources such as Nature, The Economist or The New York Times (Gale, 2017). This last resource covers several subjects, such as physical and social sciences, technology, medicine, engineering or literature, unlike Communication & Mass Media Complete, which only covers communication.

    After reading the titles, keywords and abstracts of the 521 articles selected from these two databases, some articles did not refer to data, journalism and journalist issues although they included the terms. Some articles used the word “data” referring to the data used for the research (a survey, for example) or as a synonym of information. Other articles used the words “journalism” or “journalist” referring to the author but not the content of the article. Furthermore, some articles were duplicates. After excluding all these cases, the final sample related to the use of data in journalism and journalistic practice comprised 69 articles (see Figure 1).

    These 69 articles were included in an Excel database. For each article, the registered variables were title, authors, name of journal, year, other information about the article (issue, volume, pages), a relevant sentence from the abstract, and the keywords indicated in the search tool (from the authors or the database). In addition, we gave them a category that identified an area relating data to journalism/journalist, taking into account the ones previously highlighted and listed below: