Data Journalist Playbook Table of Contents: 1. Company Description 2. Hiring a Data Journalist 3. The Interview Process 4. What is a Data Journalist? Getting Started 5. Learning the History of Data Journalism: The First Quarter 6. Taking it a Step Further: Key Skills for a Data Journalist 7. Appendix 1 i. The Final Posted Job Description ii. Content Marketing Companies to Reference iii. Interview Timeline iv. Pitch Challenge During Interview Process v. Pitch Evaluation Criteria vi. First Month Music Data Journalist Goals 8. Appendix 2 Resources for a Data Journalist
Chapter 1. Company Description What is Next Big Sound? For three years, the Next Big Sound team has been gathering data on hundreds of thousands of artists in an effort to provide the music industry with a revolutionary method of tracking progress, fan base and growth. Between the information we can glean from social media websites such as Twitter and Facebook, streaming services like Spotify, Rdio and Pandora, radio airplay as well as proprietary data on both physical and digital sales, we are in the position to offer a unique product to customers, whether they be major record labels, agents or the artists themselves. Linking this data to events or particular marketing approaches such as live Twitter interviews, the launch of a new album, or an active presence in social networks, allows the industry to track how certain strategies impact the artist and their standing, what garners the most positive fan reactions, what social networking efforts translate into sales and much, much more. With this knowledge in hand, they are able to plan better strategies, and understand the dos and don’ts for artist trying to make it in an ever-‐changing music industry. It should come as a surprise to no one that the world is becoming increasingly interconnected through social networks, and the same is true for music fans. A fan in Japan can discuss the merits of their favorite artist with their counterpart in Dubai in real-‐time (time difference notwithstanding). In addition, this coming year will mark the first in which digital downloads surpass physical sales. Music is more accessible to the masses and DIY artists are showing time and time again that it is possible to make it on
your own by taking advantage of these networks -‐ crowdsourcing funds for tours, self-‐publishing their music and videos and connecting with fans. Next Big Sound will continue to grow over the coming years, we are tirelessly working to formulate the algorithms that will allow us answer some of the biggest questions in the music industry today.
Chapter 2. Hiring A Data Journalist Alex White, CEO, explains the rationale behind the hire. As a company dedicated to providing analytics to the music business, a large part of Next Big Sound’s job is to educate music industry professionals on not only the possibilities for concrete uses of data in their day to day roles, as well as the growing importance of this. Two years in we were seeing great traction in our enterprise sales, a changing cast of competition, and a wildly shifting industry transitioning from the old way of doing things to a less certain digital frontier. Increasingly, our customers, and potential customers, were turning to us with questions of how to navigate the modern music industry. We learned early on that the only scalable way to help our customers and the industry do their jobs better was to build great software and publish our thoughts on the macro state of things that we were uniquely privy to given our worldwide data set. When we first started the company I was responsible for producing this content for marketing and awareness of the company. This responsibility passed to our first head of customer support but her bandwidth for producing regular content quickly shrank as customers came on board and demand for the product and support rose. As a board and management team we quickly realized that we needed to hire someone dedicated to helping us own mindshare around data and the music industry. This competency on staff would accelerate our march to become the music industry standard, which, in turn, allows us to build a large and lucrative business. We debated between hiring a business analyst or a data journalist for this role but decided that communicating the analysis around the data was as important as the analysis itself. The decision to hire the world’s first music data journalist was made. But what was the best way to find and screen candidates? We’d hired salesmen, software
engineers, and designers but no one in the world had ever hired a music data journalist before. Unfortunately there was no template to follow. I worked with Antony Bruno, a former writer for Billboard, to help craft the role, screening process, and evaluate candidates. We narrowed down the ideal skillset to four key components: 1. Storytelling: able to take data and extract and produce insightful narrative 2. Analytical: comfortable in Excel and able to proof underlying math and statistics before releasing to customers. 3. Design sense: able to take findings and compile professional looking blog posts, presentation decks, and PDFs 4. Music industry knowledge: experience in the music industry to help inform reports that would be interesting to our customers *Note: we had programming ability but decided to remove this and add it as a bonus in the background section. We decided to follow our typical hiring process of quick phone screens of promising resumes, 48-‐hour challenges, in-‐person interviews, and then a job offer to the top candidate. We were quickly overwhelmed with hundreds of resumes and decided that instead of the phone screen, we would solicit story pitches from the most promising candidates, narrow that pool down to the best 5-‐10 and then surprise them by having them have to turn that into a full story with data from Next Big Sound. The purpose of this exercise was to mimic the actual work, deadlines and short notice included, as much as possible. For a complete view of the job description we used, interview timeline we stuck to, pitch challenge we gave candidates, and evaluation criteria we used, please see the appendix.
Chapter 3. The Interview Process Getting hired to be Next Big Sound’s Resident Data Journalist From the moment I spotted the opening at Next Big Sound for a music data journalist, I was intrigued. At first I wondered what exactly a music data journalist does, then decided it was most certainly something I would like to find out. I had heard about Next Big Sound through working as an assignment editor on a book entitled The Human Face of Big Data, and the name came up in almost every pitch meeting with everyone excited about what the boys in Boulder were doing with music. The application process for the data journalist position started with a simple resume submission. Fortunately, it seemed my experience, education and interests up to this point suited what they were looking for in order to fill this very new role. My parents have sat through endless hours of screechy violin recitals since childhood (thankfully I am told they got less screechy with time), I ran a radio show in college, and after completing my masters in journalism at NYU a few months earlier, I had started working on a project revolving around Big Data. Not long after submitting my resume, I received an email from Alex White, CEO and co-‐founder of Next Big Sound, asking me to please submit a pitch for what type of story I would write. Those invited to pitch were free to come up with whatever idea they saw fit, but the proposed article would have to fall within one of five categories. Applying data to prove or disprove a theory on a current music industry topic, providing data-‐related insights into a topical news item, analyzing the impact of a newsworthy event on a given artist or artists, identifying trends in NBS data and surface lessons that can be applied industry-‐wide, and last but not least – case studies. The idea I came up with focused on the upcoming Coachella Music and Arts Festival in California. I wondered whether it would be possible to determine who was making the
biggest splash after performing -‐ the industry veterans like Radiohead, newcomers like Azealia Banks, or the reunion acts that the festival is well-‐known for bringing together, such as Pulp. If they liked my pitch, I would be asked to write the story, and whether it was published or not, receive $250 in compensation as well as lifetime access to the Next Big Sound platform. Not bad if you ask me – I would have done it for free. However, the idea of being compensated for my ideas and work, which is not a typical phenomenon in interview processes, is something that stuck out in my mind as a sign of appreciation and highly motivational, and also was my first clue to the management style of this company: happy employees are happy to go that extra mile for you. So, after biting my nails for a few days, I was elated to receive an email saying I had made it to the next round and that the deadline would be in two days. The piece was to be no more than 750 words, and if it was not chosen I would be free to shop it for publication elsewhere. With access to the Next Big Sound data, I set about putting together an article on my pitch. While I had to fiddle around a bit before knowing exactly how to use it, I found the platform easy to understand and quickly started to gather the information I would need in order to compare these categories of performers. I delivered the piece before deadline and set about anxiously waiting to hear. The next day, I received an email from Alex, asking me to meet with him for an interview the following week. We met and had a 45-‐minute conversation about my article, previous work, what I thought the position would and should entail, as well as what his vision was for the role of a data journalist with Next Big Sound. Following our chat, I was introduced to Yu-‐Ting Lin, the VP of Finance and Operations, and launched into what would be the most intense interview I have ever participated in. We spoke for close to two hours in a very casual manner, about anything but the standard, dull interview topics, and I found myself sharing more-‐than-‐a-‐lot about who I am.
As I was leaving the office that day, it occurred to me that this was an extremely clever way to vet future employees. I have often discussed with my peers the lack of value there is in the standard interview. Asking age-‐old questions and receiving practiced responses in return, gives you very little insight into the true personality and motivation of the candidate you are considering for hire. A few days later, I was asked to submit my references and Alex and I scheduled a follow-‐up call. In this conversation, he explained that because we would be entering unchartered territory, a large part of my role as a data journalist would be to analyze my progress, see what works and what doesn’t, and from there determine what the next step should be. In that vein, he asked me to analyze the interview process itself. I gave him my thoughts on the process, and added that I felt it might have been helpful for them to pose a question about the data to candidates that had a pre-‐determined answer, in order to truly test their comprehension. I was very excited to receive a phone call to let me know that I had been chosen for the position. Given the timing of the hire, in the midst of the company move to New York, it would it take almost a month for me to start work. After receiving an onboarding document, with ideas for what I could expect in the first few months on the job, I took advantage of that time to start building my editorial calendar. I was also happy to have the chance to fly out to Boulder, to see where it all began, while Alex packed up the last of the monitors and shipped them to New York. And once we got to the office on June 1st, I hit the ground running.
Chapter 4. What is a Data Journalist? Getting Started What? I’ve been here a month already? This afternoon marks the end of my first 30 days with Next Big Sound, and while figuring out exactly what my role is here will be an on-‐going process for some time, a lot has already happened. All newly hired engineers at Next Big Sound are asked to push out code on their very first day at work. This, as Alex explained to me, helps them get over the fear of that initial step and jump right into the fray. The data journalist equivalent of this would be to produce a post for the blog on my first day on the job. At first this seemed like a tall order, but I quickly narrowed down my ideas and got started. Writing a piece and publishing it on day one, while somewhat nerve-‐wracking, was a great way to grab hold of the blog, gain confidence in my abilities to do this job, and know that I was able to create strong narrative content based on the wealth of data we have at our fingertips. The opportunities for the type of questions I can pose, the articles I can write and for what I can learn about the music industry are seemingly endless. Alex and I sat down together and decided that for now, two blog posts a week would be a good amount to publish. In an effort to keep it consistent for our readers, I regularly post these on Tuesdays and Thursdays. Because the blog has not been regularly maintained before I came onboard, what type of content will be the most successful and interesting to our customers is still something that we are figuring out, and the best way to do that is by trying and failing. So far I have written about emerging bands, music trends from the data perspective, bands that are successful in implementing various strategies be it crowd-‐sourcing funds for touring or social network tactics, what genres are most popular at festivals, and more.
Alex and I meet on a weekly basis and discuss our progress, which I keep track of in an Excel file. Monday mornings I pull a community report with key metrics on Next Big Sound from our Premier platform, and in addition keep track of our Tumblr followers, which blog posts I have written in the past week, as well as any coverage there has been of Next Big Sound in the press. We look at any spikes or declines in our numbers and discuss what may or may not be the cause of them, attempt to analyze which type of posts are doing the best, and plan for the coming week. Another aspect of the position is an editorial contact list I am putting together. This is a list of editors and journalists at various publications, who I will reach out to and establish relationships with. The purpose of these relationships is to have an outlet to which I can supply with our data and trends that we spot, in return for coverage of Next Big Sound. Each article that mentions our data as a valuable source of information helps build our reputation as a company that is integral to the future of the music industry. Set for release in early September is the re-‐launch of our platform. As part of this, I have been working with our lead designer Andrew Cohen, and head of product and co-‐founder David Hoffman, on the redesign of the Next Big Sound blog. We have put together a page that resembles more of magazine layout with content divided into sections, a heavy focus on images, includes an NBS Top 15 playlist, and most importantly will allow us present the type of content we are able to create in a more professional manner. In addition to this, I handle our social media accounts, from Twitter to Facebook, to Google Plus. I regularly update these with a variety of content, ranging from interesting graphs I find with our platform, links to artists that are displaying surprising developments, engaging questions for our audience and more. Again, figuring out what are the most successful posts is an ongoing process, which I track through the available analytics.
In order to create the most compelling content possible, it is essential that I spend a good amount of hours each week trolling both physical and digital magazines for industry developments. I keep up to date on not only music publications such as Rolling Stone, Billboard Magazine, Pitchfork and Spin, but also more tech-‐oriented outlets such as Gizmodo, TechCrunch and Digital Music News. I cannot maintain a blog without employing a voice of authority. As the second month begins, I am beginning to understand that the position of data journalist is comprehensive. It will require me to not only be on the ball in terms of breaking stories, but also to maintain a long-‐term perspective on developing the outward face of this company. By keeping careful track of my progress and carefully scheduling my workdays, I am able to manage these responsibilities.
Chapter 5. Learning the History of Data Journalism: The First Quarter Three months in, and it is hard to imagine where all that time went, how little I have slept, and just how much I have learned. Summer being one of the busiest seasons in the industry, with festivals staged almost every weekend, outdoor concerts by the bushel, tours across the country and story ideas for a new data journalist popping up every minute. Since taking on the role as resident data journalist, I have been fielding a lot of questions about what exactly this means. The concept of including this type of information as a basis for articles is anything but new, a widely cited example is the use of educational data for an article in the very first issue of the Guardian in 1821. What has revolutionized this field in recent years is the amount of data available, and the speed with which this data is generated and delivered. At Next Big Sound we are gathering an average of 175 million data points each day. There are several great examples of journalists who use data heavily in their work. Some of the best stories to hit the press this past year are articles based on data findings, such as the investigative series on horse racing in the New York Times entitled Death and Disarray at America’s Racetracks. Data journalism can also come in different formats, for instance, the News Application team at the Chicago Tribune consists of a group of programmers embedded in the newsroom, assisting journalists in uncovering data and creating visualizations. The magnitude of information now being gathered and stored within most fields, from healthcare to consumer behavior to the various social sciences, serves as an invaluable resource to those writing the news. At this point, I am only just beginning to comprehend the endless opportunities for what kind of stories I can write based on this massive amount of data, and just how important it is to fuse this type of information into an industry that can be reluctant to the idea of
change, but is rapidly changing nonetheless. I feel more in the loop when it comes to the industry, understand the ins-‐and-‐outs of our platform, and am able to quickly determine what stories our audience will respond to and how. At Next Big Sound, I have at my fingertips a platform that allows me to easily graph information in order to see correlations, as well as the ability to pull overview reports of relevant data. Telling great stories then simply becomes a matter of figuring out the right questions to ask of the data and combining this with relevant reported content. Working as an embedded data journalist with a company can of course be challenging without an editorial team around me to bounce off ideas. However, I often use my colleagues as a sounding board and given all the stories the data has to tell, have yet to come up empty-‐handed when deadlines roll around. Another aspect of the position that has risen in importance in the past few months has been ensuring the distribution of our content, through more than one channel. I find myself being interviewed about this new type of role, speaking on panels about the future of the industry, building individual relationships with editors and journalists that are interested in applying new data to the questions they are posing. In addition to this I am now working with several online publications to further syndicate the content of our blog, among them the MTV O Music Awards blog, Sidewinder.fm, Hypebot and others. As we continue to grow, I plan to cultivate more and more of these relationships in order to ensure that our content and mindshare around the data we have available at Next Big Sound is widespread and becomes part of the daily conversation in the music industry. From time to time, I will describe my job as basically “de-‐nerdifying” Next Big Sound. Working on such a technical level, it can be a challenge for colleagues to communicate in a simple terms what they are doing. Here is where my listening, comprehension and communication skills come in handy. I take the complicated data science projects that
they are working on, such as how the concept Granger Causality can be used to calculate the causation between social media metrics and record sales, and explain the significance of this to an industry that can be hesitant of using data.
Chapter 6. Taking it a Step Further: Key Skills for a Data Journalist As I am nearing my first six months with Next Big Sound, I have learned a great deal more about the basic skills that will help me do my job better, in terms of dealing with the process and presentation of data. Presentation When you are presenting data in graphical form, there are several considerations you must make. Understanding the numbers is of course imperative, but presenting the numbers in a manner your audience will easily comprehend is the most pressing challenge. The advantage of working for Next Big Sound is that I very rarely have to think about how to present the data; it is easily done for me through a platform that is already carefully crafted and considered. But for the typical data journalist, knowing how to present statistical information in a way that intrigues and involves the reader can be a challenge. Taking care to not obfuscate the data, and creating charts that are easily readable. One of the greatest resources I found in learning how to deal with data was the Wall Street Journal Guide to Information Graphics, written by Dona M. Wong. This book gives you a comprehensive insight into the basics of presenting statistical information, and the dos and don’ts of charting data. Extremely useful, this guide is an overview of key rules that will help any data journalist understand how to present data in a manner that is not only valuable, but also completely accurate. I don’t for a minute wish to imply that I now know everything about graphical design or charts, but I have learned the basics of how to create a basic visual representation of data from this book, and believe it to be integral to explaining this phenomenon we call Big Data.
Learning to query with R Another step in becoming integrated into this company, was realizing I knew next to nothing about how to deal with information stored in a database. A database is a central collection of information, organized in tables, which when relevant and utilized in the correct manner, will be a source of answers, to any question you might have. Working with a team of engineers, data scientists, and designers who understood programming like the back of their hand, it became imminently obvious to me that the ability to extract information from whatever database I had at hand, whether it that of that Next Big Sound, or any other, was integral to independently mastering this role. In this vein I decided to delve into learning how to query. Now, the differences in programming language can be somewhat complicated for those of us who aren’t engineers. There are several ways in which you can interact with a database, whether it be languages such as Java, Python, R, or whatever else these guys (who are definitely smarter than me) come up with. But a basic requirement, I believe, for a data journalist, is to be able to extract the relevant information you need for a specific story from it. The initial step is to learn Sequence Query Language (SQL), a standard vernacular for computer programming which allows you pull information from a database. For instance, if you would like to get a list of all artists whose name starts with the letter B, or who have between 5000 and 50,000 fans on Facebook from the Next Big Sound database, you would need to know the basic commands of SQL, such as SELECT, AND/OR etc. There are several free online resources where you can learn the basics. Taking this a step further, you might like to narrow down this query to a more specific question, such as, how many of these artists gained a certain amount of followers, within a certain time period, and how rapidly did that growth occur. Our data scientists rely heavily on R, which is the language I am in the process of learning. Using this I can
ask more complicated questions of the data, take a stab at various graph formats in order to see what might work best, and eliminate data that for various reasons might not be relevant. With the ability to query, the opportunities for what you can learn from a collection of numbers starts to become not only unimaginable, but also overwhelming. Over the coming months, I will continue to study the various programming languages, in hopes of eventually mastering them. As data journalists, we are opening a whole new world. In terms of what we can do, what we can learn, and what we can explain to our audience. And as this world of data grows larger, faster and ever more unmanageable, it is our job to understand, even though that means stepping outside our comfort zone of writing, recording and editing information. You may never have thought that as a journalist that you would have to learn how to code, but now that is becoming a basic requirement of telling the narrative of Big Data.
Appendix 1 Actual NBS Data Journalist Job Description Good References to Study Interview Timeline Pitch Challenge Interview Process Pitch Evaluation Criteria First Month Music Data Journalist Goals Job Description (Note how NBS still asked for business analyst) We are a two-‐year-‐old, venture backed, tech startup focused on measuring the music industry, both online and off. We collect engagement metrics from all the largest social media sites on a daily basis, we catalogue events like releases and concerts, and we import historical archives of radio and sales data from our major customers. We are looking for a data journalist/business analyst to join our growing team of highly competent engineers, designers, salesmen and product developers. Ideal skills in order of importance: Storytelling: able to extract data and produce insightful narrative. Analytical: comfortable in Excel and able to proof underlying math and statistics before releasing to customers. Design sense: able to take findings and compile professional looking blog posts, presentation decks, and pdfs. Music industry knowledge: experience in the music industry to help inform reports that would be interesting to our customers. Main responsibilities: • Create constant stream of content and analysis by writing timely articles, industry-‐wide macro reports, event response measurement (Grammys etc.) • Create compelling examples for marketing collateral, key live conference presentations, and sales pitches • Field ad hoc request from major users and individuals in support of NBS account managers • Build case studies and best practices • Background: • Journalism background (degree or newsroom experience)
• Passion for music and the music industry • Bonus: working SQL or programming knowledge Good References to Study http://contently.com/blog/ http://blog.okcupid.com http://blog.redfin.com http://blog.runkeeper.com/ http://37signals.com/svn/ http://blog.birchbox.com/ http://www.seomoz.org/blog http://www.etsy.com/blog/en/ http://blog.uber.com/ Interview Timeline Week 1: - Post Job description posted for Data Journalist Week 5: -‐ Narrow down resumes to the best 10-‐20 -‐ Finalize first-‐cut names -‐ Send first-‐cut candidates a Pitch Challenge due in 48 hours. Candidates should submit a pitch consisting of: -‐ Summary of a story idea they’d write today (based on topical news and/or trends observed in your data) no longer than 750 words -‐ Why that story idea should be considered -‐ Target audience and desired audience takeaway (Note: in pitch challenge, communicate when selected pitches will be notified) Week 6: -‐ Narrow pitches down to the best 5
-‐ Notify the selected pitches to write their stories in 48 hours Week 7-‐8: - Finalize the best three candidates and conduct in-‐person interviews Week 9: - Send offer letter to the best candidate Pitch Challenge from Interview Process (template) Thank you for your interest in the Data Journalist position here at <Your Company Name>. We’re pleased to inform you that you’ve made the “first cut” of applicants, but we want to get to know you a bit better before moving to the next stage. We’d like you to pitch us on a story idea for the type of piece you would write if this was your job. What we’re looking for here is your ability to identify compelling story ideas and angles that would catch the eye of industry executives and the press. Pitches can run the gamut from: -‐ Applying data to prove/disprove a theory on a current industry debate -‐ Provide data-‐related insights into a topical news item -‐ Analyze the impact of a newsworthy event -‐ Identify trends in our data and surface lessons that can be applied industry-‐wide -‐ Case studies We will be making our second cut based on the strength of these pitches, so make ‘em count. Pitches will be due by close of business <date>. Pitch Evaluation Criteria __ Identifies a timely, relevant topic __ Applies data properly and relevantly __ Demonstrates understanding of your business / industry __ Demonstrates understanding of data analytics __ Demonstrates an understanding of news hooks/angles __ Demonstrates creative, critical, and out-‐of-‐the-‐box thinking