Project: Data-Expanded Story


None. Delivery won’t take place over a single-email
2017-12-12 23:59
Use everything you’ve learned in this class to greatly expand on the research, depth, and data analysis of a story that you’re already producing for another class or publication


Past versions of this course have asked for students to produce a “data story”. As most students in this class are already very booked with their beats and stories for COMM 275 & 279. The requirement to do a “data story” has usually led students to do an entirely different topic than what they’ve covered. Which is a messy thing, even before considering any lack of confidence they may have with data in general.

That’s not a good experience for students or for me. And it’s also quite counter-productive with how I see data being inseparable from strong journalism.

I’m going to iron out the details of this over the next couple of weeks. But what I have in mind is that at the end of the quarter, the j-school students are focused on their features for 275 and 279. I want students to use the data and research skills they’ve learned in this class to greatly expand the scope of their end-of-quarter story for COMM 275.

I’m specifically not requiring that you make a visualization, or some other kind of data-doo-dad. Though making visualizations, if I teach this course well, should be something that is a natural and easy enhancement to your story, not a chore.

In general, I would expect the data-enhanced version of your story to greatly expand its scope, whether that’s raw word count, or more ideally, deeper coverage.

At the very least, I would expect your story to utilize at least 30 to 50 different interview, documentary , and data stories. Not as an arbitrary quota of things to include, but because you know enough about your story to have that depth.

Last year, there were a couple of assignments focused on using LexisNexis to as a way to explore the history of a topic. For virtually any given event, something similar has happened in the past. Or elsewhere. And what happened then gives context to the story today, even if it’s as simple as “This was worse/better than X years ago.”

I’ll flesh out details in the coming weeks. The main upshot is that whatever your beat is in 275+279, you’ll be thinking about it in this class as well.


Here’s a story by former student Peter Arcuni (class of 2017) to the Peninsula Press that came from his graduate thesis project:

Besides being a great example of how a story can be told using a variety of skills – writing, photography, audio, data analysis, and visualization – Peter’s story is just a great and important read.

I don’t/can’t expect the final project/story for this class to have the same amount of depth as a thesis project. But here are a few things about Peter’s story that I definitely do think you should be able to do for your story for this class:

The fate of Box City and San Francisco’s enduring homeless camp epidemic

  • Interviews, data, and records from several agencies and organization:

  • Analysis of records/data from 2+ databases:
    • “In 2016, complaints about the encampments logged by the public to San Francisco’s 311 customer service number swelled to upwards of 2,500 per month.”
    • “A database kept by the Department of Public Works to monitor encampments for its cleanup crews reveals that the number of active camps has remained static, hovering around 70, since the start of 2017.”
  • 2 straightforward visualizations:
    • Line chart showing the aggregate county of 311 complaints by month
    • Carto “bubble” map showing where the roughly 70 officially known encampments are.

Note that you don’t have to do any visualization for this story to meet the class requirements. I’m confident that you will because if you meet the other requirements for sourcing and research and analysis, you’ll have done all the work and reflection needed to create visualizations as appropriate.

What was time consuming about this story (besides its editing and production) is the in-person reporting that Peter had to do. The photos of the encampments that have his byline meant that he actually had to go out and visit these camps. Likely, it took several visits to these camps with his camera and notebook.

Peter’s story references a decent number of agencies and groups, but his story focuses on a 44-year-old Box City resident named Roland. Roland might likely be one of several people that Peter interviewed and followed-up with. The best anecdote and subjects are rarely the first anecdote/subjects you come across.

For this class, you’ll likely be limited in the amount of in-person “shoe-leather” reporting you can do (I mean, you’re already doing it for your 2 other classes).

But with research, curiosity, and knowledge of research tools and services, you can report and include facts across dimensions.

Take this fact from Peter’s story:

Over 4,300 homeless San Franciscans remain unsheltered each night, according to the latest estimates.

If you were to write that factoid in your own story, these are all the questions I would challenge you with. It’s not that you have to answer these meta-questions in the story, but you do have to know how to answer them:

Easy questions

Peter’s story reports on a number, an aggregate count, made this year by an official agency.

It doesn’t matter that this number is specifically about homeless population, or that it involves SF. The following questions are universal data quesations that you should always seek to answer:

  • What was last year’s count?
  • Is this year’s count more/less than last year? By how much?
  • What were the counts in the last 5, 10 years?
  • How does this year’s count compared to the 5/10 year average
  • SF is not the only city in the Bay Area with difficult housing problems. What is the homeless count for San Jose?
  • What is it for other Californian regions, e.g. Los Angeles, and how does SF compare?
  • If SF experienced a steep rise, then what are the historical numbers for those other entities. And do they also indicate a steep rise in homelessness in those respective jurisdictions?

More tenuous questions

Even questions that demand longer, more detailed answers, are pretty straightforward:

For the homeless county, Peter links to this infographic:

– which sources: “Applied Survey Research. (2017). San Francisco County Homeless Count & Survey. Watsonville, CA.”“

Without getting into the specific details of what a homeless count is, these questions apply in general:

  • Why should we trust the agency/agencies who came up with this homelessness count?
  • What is their methodology, i.e. how and how often are they counting, because homeless people don’t have static addresses?
  • Who else uses the same methodology (different cities, counties, countries?)
  • Who else uses their methodology? Do the neighboring counties? The region? The state? If not many others, then why not?

If you can answer the above questions, or if you know where to go to find those answers, then you’ll have the research/data part of this story done pretty easily.

Finding humans

But what about the humans. Peter’s story has a lot of useful facts and context. But at the heart of his story was 44-year-old Roland. Finding someone like Roland is non-trivial, and unlike most data work, non-deterministic. Even if we asked Peter how he found Roland and came to realize Roland would be the emotional center of this story, that advice doesn’t directly help us.

Maybe Peter lives in SF and knows the city well, including where encampments take place. Maybe he has friends in an homeless advocacy organization, or in SF law/code enforcement. Maybe Roland is a friend of a friend of a friend. These are all unpredictable things, and things that may be completely inaccessible to you as a student reporter.

So start with what you know how to predictably and mechanically use: if SF keeps a database of known and active homeless encampments, that is, that means that at the very least it can serve as a list of places that even someone new to the city could use to locate and visit the biggest (or even the most problematic) camps. If your story needs voices from homeless people, then the most important and obvious measure you can take is: go where homeless people are. The encampment database fan be used to provide an analysis. But it can also just be used as information – publicly-accessible expertise that you use to quickly get up to speed.