The state of local data journalism

By • The Data Journalism Podcast in Zoom

Recording

Listen on the original podcast site

Show the timestamped transcript
  1. [MUSIC PLAYING]
  2. Hello, and welcome to the Data Journalism Podcast.
  3. My name is Simon Rogers.
  4. I am a Data Journalist, speaker, and teacher, and data
  5. editor at Google.
  6. And my name is Alberto Cairo.
  7. I am a professor of Visualization
  8. at the University of Miami, an Infographics designer,
  9. and journalist, and also a book author.
  10. We love using data to tell stories.
  11. And the music you can hear is the sound of data,
  12. made with two-tone, an app that turns numbers into tunes.
  13. There's also the sound of football, American football,
  14. that is, with a changing football school since 1920,
  15. care of one of our guests today.
  16. And this is the Data Journalism Podcast
  17. that dissects the latest trends in data journalism
  18. around the world.
  19. In each episode, we will explore the latest in data journalism,
  20. and we will chat with some of the world's top data journalists.
  21. You'll get to find out how they do what they do.
  22. So subscribe at datajournalismpodcast.com
  23. to see how data is changing the world of journalism forever.
  24. Hi, Simon.
  25. Hey, Alberto.
  26. How are you doing?
  27. So good.
  28. I'm really thrilled to have the chat that we had this week.
  29. It was a lot of fun.
  30. It was a very interesting one, right?
  31. Can you tell us a little bit about our guests?
  32. Yeah, so we have two doyen of local data journalism.
  33. We've got Ben Welsh from the LA Times, who
  34. has been just one of the first people out there
  35. to really see the power of using data to tell stories.
  36. The Los Angeles Times have one of the first great data
  37. journalism sections, and he has forgotten more than we will ever
  38. learn about how data can just improve local reporting.
  39. And we also are lucky enough to have Mary Jo Webster, who
  40. is a great local reporter who's just taught herself
  41. how to work in this area and really created
  42. some interesting work.
  43. It was a very interesting conversation with both of them,
  44. and we got into sort of how they got started,
  45. how they discovered the tools that they use nowadays,
  46. beginning with Excel.
  47. They tell a couple of funny stories
  48. about how they discovered Excel and then later more
  49. sophisticated data analysis tools.
  50. I think that it was a very enlightening conversation,
  51. and listeners will enjoy it a lot.
  52. We should do an episode one day about Excel
  53. and the power of this tool in getting people
  54. into data journalism.
  55. Love it or hate it.
  56. It's just become this kind of like entry-level drug
  57. for people who are getting into data and reporting.
  58. I mean, it's super useful.
  59. You can do amazing things in Excel, right?
  60. I mean, both Ben and Mary Jo, they
  61. mentioned AR, the programming language, and Python,
  62. and many other tools.
  63. But both of them say, you know, if you
  64. can get away with doing this very simple thing with Excel,
  65. why don't do it with Excel?
  66. Or if you are getting started in this field,
  67. it's better to begin with the tools that you already know
  68. and many people already know that it's Excel, right?
  69. Yeah, and one of the things we wanted to do with this episode
  70. was really getting some of the nitty-gritty.
  71. So we have a rapid-fire series of questions at the end,
  72. really about the kind of people's top tools,
  73. some of their foibles, some of the stories they care about.
  74. It's a lot of fun.
  75. And I would really advise sticking around to that,
  76. because you'll learn a lot about how each of them works.
  77. It was an excellent episode, I think.
  78. So let's get into it.
  79. Let's get into it.
  80. [MUSIC PLAYING]
  81. [MUSIC PLAYING]
  82. Hi, I'm Mary Jo Webster.
  83. I'm the data editor at the Star Tribune in Minneapolis.
  84. And hello, I'm Ben Welsh.
  85. I live in Los Angeles, where I work on a project
  86. at Stanford University called Big Local News
  87. and at the Los Angeles Times.
  88. Thank you so much for joining us, Mary Jo and Ben.
  89. We're really happy to have you both here today to talk really
  90. about local data journalism and how it's
  91. growing around the country, and I think around the world
  92. as well.
  93. But I would love you to talk a little bit first about how
  94. you got into this world.
  95. Now, Ben, when I was at The Guardian back in the day,
  96. when we were setting up the data blog,
  97. we were definitely looking at the work
  98. that you guys were doing.
  99. And I feel like you were actually really
  100. one of the innovators in this field.
  101. So talk to me a little bit about how
  102. you got into data journalism and the kind of work that you do.
  103. Well, it definitely wasn't the plan.
  104. It wasn't anything that happened on purpose,
  105. but I'm really glad it did.
  106. I was a pretty listless undergraduate in Chicago,
  107. Illinois at DePaul University and kind of unclear
  108. of where I was going with my life and career,
  109. like perhaps many undergraduates that Alberto
  110. encounters from time to time, and maybe himself as well.
  111. And I won't bore you with the whole long story,
  112. but through sort of an unusual series of events,
  113. I became the assistant to two local television journalists
  114. in Chicago, who were both excellent practitioners.
  115. And I became what was known at that time
  116. as the foot man, which was a very glorified term
  117. for the intern or the helper around the office.
  118. And so as they prepare a weekly newspaper column
  119. and television piece, I would do research and scheduling
  120. and a little bit of television editing too.
  121. And in that, I kind of caught the journalism bug.
  122. And one of the early stories we worked on
  123. was about corruption in a Chicago suburb.
  124. Locals could probably guess which ones.
  125. And we did my first public records request
  126. for legal fees from a sort of local heavy
  127. who was milking the city out of money.
  128. And it came as a really, really long printout.
  129. Remember when printouts used to have the perforated tears
  130. along each side and a single print over 20, 30 pages
  131. will be connected through these long strips in the side
  132. that you have to rip off, right?
  133. And we got that back, this epic printout
  134. of all the legal fees.
  135. And we thought, well, what are we gonna do with this?
  136. How are we gonna turn this into a story?
  137. And I don't even know from where,
  138. but I got the idea of just typing it into a spreadsheet.
  139. I think it was probably the first spreadsheet
  140. of my entire life.
  141. And I spent having nothing better to do.
  142. I spent a good chunk of time just hammering all that in
  143. and we added it up.
  144. And at the end, that number got read out on the local news.
  145. The anchor even climbed up a step ladder
  146. and dropped the sort of accordion of paper live on camera,
  147. which was kind of a funny little gag.
  148. And I kind of caught the bug.
  149. You know what I mean?
  150. And within that, I saw that through technology and data,
  151. it was kind of like a way I could help.
  152. It's something I could chip in and do
  153. as a place for me to contribute.
  154. And it also was a gateway to doing more ambitious journalism
  155. than just the average run of the mill story.
  156. We did a story nobody else was doing
  157. that was more ambitious and analytical,
  158. and that was exciting.
  159. And so I kind of just got into it there.
  160. This is now 20 years ago, I guess,
  161. or more than that actually.
  162. And I just followed that and followed that.
  163. And without boring you with my whole life story,
  164. I mean, I kind of ended up on this podcast today.
  165. - Mary Jo, tell us about how you caught the bug.
  166. - Mine is very much a local news-based situation.
  167. I was, I don't know, three or four years into my career
  168. working as a reporter in Oshkosh, Wisconsin,
  169. a very small daily paper.
  170. I was tasked with covering the police and sheriff's department
  171. both on a kind of a policy standpoint,
  172. but also, you know, I went out to the crime scenes
  173. and the fires and whatnot.
  174. But I was covering the city budget
  175. and it looked like the police department
  176. was growing its budget dramatically one year to the next.
  177. And I was sitting there with a calculator
  178. trying to calculate the percent change
  179. between the two years for all of the departments.
  180. And a colleague of mine came over and said,
  181. do you know what Excel is?
  182. And I was like, no, no idea.
  183. He opened it up.
  184. He showed me how to do a simple percent change formula.
  185. My eyes were bugging out.
  186. I was like, wow, this is so cool.
  187. Where did you learn this?
  188. You see, he told me I went to this IRE&NICAR bootcamp
  189. and they teach you all this stuff called,
  190. back then it was called computer-assisted reporting.
  191. And I was immediately hooked
  192. and I decided I had to keep going.
  193. And sometime later that year,
  194. I was able to attend a training that Sarah Cohen did.
  195. She was the IRE trainer at the time.
  196. And I remember sitting in this room.
  197. I don't think it was a hands-on training.
  198. My memory serves.
  199. But I just remember being enthralled
  200. by everything she was saying
  201. and mind being blown by this entirely different way
  202. of reporting that I had not learned in college
  203. that my editors had not taught me in my early years.
  204. And it was a complete turning point for me.
  205. - That's amazing.
  206. Both historical, welcome to the podcast in the first place.
  207. So thank you for being here.
  208. But the stories that you are telling
  209. are similar to some of the stories
  210. that I tell my students that
  211. when you see data journalism or my own field,
  212. data visualization from the outside,
  213. it looks like these very shiny, super advanced,
  214. high technology kind of endeavor.
  215. But when you see how this software is made,
  216. it's actually sometimes pretty pedestrian, right?
  217. It's like you essentially type the data
  218. into an Excel spreadsheet
  219. or you don't use machine learning algorithm.
  220. No, you just use a simple Excel pivot table, right?
  221. So I find all that absolutely fascinating.
  222. I mean, we are going to go back and forth
  223. between your professional careers,
  224. but also other sites of your professional endeavors.
  225. I'm really curious about your educational endeavors
  226. because both of you have been involved
  227. in educational initiatives, right?
  228. Can you talk a little bit about that?
  229. Both of you?
  230. - You mean what training we have taken or the teaching?
  231. - The training that you offer,
  232. because for example, Mary Jo,
  233. you have a data journalism academy, right?
  234. So both of you have been involved
  235. in these types of initiatives.
  236. Not only, I mean, you didn't take advantage
  237. as of things such as like IRE NYCAR,
  238. but you also offered to the community.
  239. - Yes, I am a huge believer in teaching.
  240. I think it has helped me
  241. as much as it helps the other people
  242. because I believe very much in the learn, do, teach approach.
  243. I found all the holes in my knowledge
  244. by having to turn around and teach it to somebody else.
  245. So I've also gotten so much back from other people.
  246. That's how I learned.
  247. So I feel the need to turn around
  248. and teach it to other people.
  249. I have taught for about 10 years as an adjunct
  250. at the University of Minnesota,
  251. a data journalism class that was fully dedicated
  252. just to data journalism.
  253. I'm currently a senior fellow
  254. with the Center for Health Journalism's
  255. Data Journalism Program,
  256. in that we do a week-long training
  257. and then we do six months of mentoring
  258. and additional training throughout that time.
  259. I'm particularly enjoying the mentoring part of that
  260. because it's another level of teaching.
  261. You're helping them get a project off the ground.
  262. You didn't just teach them how to run things in Excel
  263. and then walk away.
  264. You keep it going and help them get a story out the door.
  265. And it's quite fulfilling, actually.
  266. And then I taught at IRE and Nicar a lot
  267. and I built up so many materials
  268. that I've gradually decided I should share those materials.
  269. And at the start of the pandemic
  270. when I had a bunch of free time, I thought,
  271. you know what, it's time to turn some of these materials
  272. into videos.
  273. So I made a new website and posted a lot of stuff there,
  274. mainly Excel and other teaching things.
  275. I am doing a new thing right now
  276. that I'm actually gonna present about this
  277. at the Nicar conference this year.
  278. I've created a little data for editors training module
  279. for the assignment editors in our room.
  280. And so these are the people who work directly
  281. with the reporters, mainly in our local news section.
  282. And I've broken it down.
  283. I'm not teaching them how to use Excel
  284. or how to run data or any of that.
  285. I've broken it down into little sections
  286. of the different points in a data-driven story
  287. that the editor needs to weigh in and look at
  288. and what they should be doing at each point
  289. and what they should be looking for.
  290. So like the key ones are that brainstorming process
  291. with their reporters when they're talking
  292. about what story should I do?
  293. What should I tackle?
  294. Looking, having, I'm giving them some clues
  295. that they should be watching for
  296. that maybe there's a data opportunity here.
  297. What questions they should be asking, things like that.
  298. And then the bulletproofing process.
  299. Okay, the data has come in.
  300. We've got findings.
  301. We're showing you a bunch of charts and tables.
  302. What do you do with it?
  303. Too often the editors just look at it and go,
  304. oh, that's pretty.
  305. Oh, that's kind of interesting.
  306. And I'm trying to teach them like,
  307. how about we take it to another level
  308. and you really help bulletproof it.
  309. And then the third level is the editing the written story.
  310. One of the things I've seen a lot as editors
  311. see a paragraph full of numbers
  312. and they just glaze right over it.
  313. And they don't think about like,
  314. what are we really saying?
  315. What are the key points we're getting across?
  316. So I'm trying to teach them some editing things around,
  317. you know, what's the key finding?
  318. You know, not lumping too many numbers in a paragraph.
  319. All these kinds of things to try to help
  320. make the story more powerful.
  321. And then we're also gonna do a graphic session.
  322. My colleague at the head of our graphics department
  323. is gonna help me with that.
  324. And that'll be more about, okay, when should we use graphics?
  325. What kinds of graphics should we use?
  326. What should you as an editor be doing
  327. when we send you those proofs?
  328. Instead of just saying, oh, that looks beautiful.
  329. What else should you be doing?
  330. And I'm really anxious to see what kind of outcome that gets.
  331. If it leads to more data-driven stories on a beat basis
  332. and more daily basis or a weekly basis,
  333. not just the big enterprise kind of stories.
  334. - What about you, Ben?
  335. Can I talk a little bit about your educational efforts?
  336. - Well, I think Mary Jo said something
  337. very sort of succinct and wise there
  338. with the learn D2, learn, do, teach formulation.
  339. You know what I mean?
  340. It's a great way.
  341. It definitely benefits yourself.
  342. But I also think there's another great thing
  343. about that kind of approach or model
  344. is there's really a very sort of like stimulating
  345. and actually rewarding creative process
  346. of converting something that you've learned how to do
  347. into like a lesson or something
  348. that you can teach someone else.
  349. And it actually is like, it kind of gets the neurons firing
  350. the same way when you're like kind of innovating the code
  351. in the first place for the idea.
  352. And so for me along those lines, you know,
  353. one like sort of approach to teaching that I've enjoyed
  354. or found fruitful is taking emerging new techniques
  355. that we're developing in the newsroom
  356. that I feel like are kind of like a big deal for me
  357. or a big deal for our team,
  358. but that I don't see like being taught
  359. and then trying to convert that
  360. into like a night-card bootcamp or class
  361. to try to spread it to other people.
  362. And so, you know, examples I've done this with not alone.
  363. I always try to seek out other folks
  364. to help me like really strengthen the idea.
  365. But with others, we've created some of the earliest
  366. night-card bootcamps on how to use Jupyter Notebooks,
  367. how to use the static site frameworks
  368. that so many newsrooms use to make standalone visual stories
  369. but like nobody ever talks about, right?
  370. And this year at night-card, we're gonna have,
  371. I think the first kind of three hour bootcamp
  372. on how to automate a scraper
  373. within the GitHub actions framework.
  374. And so to me, like those three things, Jupyter Notebooks,
  375. static site page building, Git scraping
  376. are three really important like technologies
  377. that have taken off in our world
  378. in the last five or 10 years.
  379. And it's just been fun to try to like think about
  380. and pin down, well, how exactly do we use these?
  381. What are the fundamental skills?
  382. How can I pass that on to someone else?
  383. - It's really interesting to me
  384. how you're both so involved in teaching.
  385. And I wonder, does that reflect towards,
  386. I wonder if this is particularly applies to people
  387. working with data on a local level,
  388. is often people are kind of isolated.
  389. Often it's like one or two people on their own,
  390. they've got no support, there's no network.
  391. But the colleagues from other organizations
  392. are kind of struggling in the same way.
  393. So there's a kind of a need there, is that fair?
  394. - I love the psychoanalytical take here.
  395. I mean, I did grow up as a lonely, nerdy kid
  396. on a grand full road side,
  397. who didn't have a lot of friends.
  398. And so maybe this is me still trying to connect,
  399. you know what I mean, and get over.
  400. I was not crazy.
  401. - Well, I think you're right.
  402. There's so many that, you know,
  403. these days they refer to themselves as lonely coders.
  404. The majority of my 20 years in data journalism
  405. was spent in a newsroom where I didn't have support
  406. from anybody else.
  407. And so I sought it out from other people
  408. and learned from other people.
  409. And then, you know, I turned around and taught.
  410. But I think another factor is the Nykar community
  411. has always been a sharing community.
  412. If you even look at how Nykar was formed in the first place,
  413. it was really a group of investigative reporters
  414. who realized they needed to learn how to work with data
  415. and they didn't know.
  416. And a few of them learned somehow, I'm not sure how,
  417. and then they decided to share it with the others
  418. to try to pass it along.
  419. And so it was really, that's how Nykar got started.
  420. And that has been infused throughout all the years.
  421. I remember so many boot camps,
  422. and I think Ben is guilty of this,
  423. of people sitting at the bar at the end of the conference
  424. helping somebody with their data, you know, hovering over
  425. a laptop, helping someone with their data
  426. when they should be, you know, having fun
  427. and enjoying themselves.
  428. But that's how much I think our community likes to share
  429. and teach each other and it just, it's wonderful.
  430. - Yeah, and just to add to that,
  431. I mean, I think if you follow the origins back
  432. like Mary Jo was doing, I mean, data journalism's roots
  433. go a lot further than I think is commonly recognized.
  434. You know what I mean?
  435. It's really a practice that stretches back decades
  436. into the invention of the computer in the '60s and '70s.
  437. And I think is that culture of those early days
  438. is probably in like an anthropological sense,
  439. like indistinguishable from the hacker culture of MIT
  440. that's so celebrated.
  441. You know what I mean?
  442. It really was a very similar people
  443. with very similar sort of ethos of sharing
  444. that is the wellspring for a lot of what we do today.
  445. And I think that, you know, as an inheritor of that,
  446. I'm continuing to be inspired by it
  447. and I feel some obligation to uphold it, you know?
  448. - I think that we all do.
  449. We all have the sort of like feel the need
  450. to contribute back to the community that helped us
  451. at the beginning of our career, right?
  452. I have a question related to what you were both saying
  453. about, you know, growing up as a nerdy kid, right?
  454. I sort of sympathize with that because, I mean,
  455. people don't see where I'm sitting right now,
  456. but I'm sitting in a room with my kid
  457. and then surrounded by comic books, science fiction novels,
  458. you know, books about history and psychology
  459. and then some tabletop games.
  460. So the nerdery comes naturally to me.
  461. So has that improved or has that changed?
  462. It's like, are we still seeing or are you still seeing
  463. as sort of like the nerds in the newsroom
  464. or has that situation changed
  465. and data journalism, data visualization,
  466. computer assistant reporting,
  467. has it has become more widespread and more accepted
  468. or more popular within local newsrooms?
  469. What do you think?
  470. - I think it is more widely accepted
  471. as a tool that every reporter should have in their toolbox.
  472. I like to talk about data analysis
  473. as really another form of interviewing a source.
  474. That's all we're doing really.
  475. And it just requires kind of a little bit
  476. of a foreign language to do it.
  477. I also see a data journalists increasingly
  478. as the innovators in the room.
  479. In all the newsrooms I've been in
  480. for quite a few years now,
  481. the data journalists have been at the forefront
  482. of whatever new digital innovation we're gonna do.
  483. And they're the ones pushing for it and seeking it out.
  484. And in my newsroom right now,
  485. my team is the one often called upon
  486. for like, what do we do next?
  487. Where do we go?
  488. And I think that's wonderful.
  489. And I think that takes us out of the nerdy idea
  490. into kind of more of the leaders.
  491. - Yeah, and you see this in the product realm too,
  492. where a lot of people out of the data journalism world
  493. have become leaders in the next generation
  494. of digital business products
  495. their companies are putting together,
  496. which is a whole other level beyond
  497. just like what's happening in the newsroom.
  498. To be frank with you,
  499. I think in the last two or three years,
  500. we've seen a quantum leap with data journalism in newsrooms.
  501. The takeover has happened.
  502. You see it at the prestige level with those metrics.
  503. You look at the Pulitzer Prizes last year,
  504. five of the Pulitzer Prizes
  505. went to data journalism projects,
  506. including the public service highest award,
  507. which went to the New York Times COVID tracking effort, right?
  508. Which on its own was a gigantic public service
  509. and an incredible feat of like data gathering and analysis.
  510. But it also was the most popular
  511. and I'll say it most lucrative thing
  512. the New York Times has ever published.
  513. And those COVID tracking things
  514. I bet at the Minneapolis paper
  515. was the most popular thing ever published.
  516. I'll just throw that out there.
  517. Mary Jo can correct me.
  518. It's the most popular thing
  519. the Washington Post ever published.
  520. It's the most popular thing
  521. the Los Angeles Times published.
  522. And I don't mean to demean the news by saying that,
  523. but once the product you're creating
  524. is both the high prestige thing
  525. and more importantly, the economic and economic engine.
  526. You know what I mean?
  527. That's growing the digital audience
  528. of which there's a desperate need to grow.
  529. The entire way the organization looks at you changes
  530. just fundamentally.
  531. And that's happening.
  532. - Yes, the tracker we'd put up was,
  533. is hands down the historically most read page
  534. in the history of the Star Tribune.
  535. And the numbers on it are so astronomical
  536. that I don't envision it being that record being broken.
  537. So it would take something pretty massive,
  538. I think to break it.
  539. But the other interesting thing we saw
  540. is that it was also the most popular page
  541. among our digital readers who are not subscribers,
  542. but they come to us pretty regularly
  543. and we would like them to be subscribers.
  544. We call them in tenders.
  545. And it was the most popular page among them.
  546. And now we have conversations going around
  547. what else like that, maybe not data,
  548. but more in the service journalism realm,
  549. but data is one of the things you can do
  550. for service journalism.
  551. What can we do to help get more of those people
  552. to become subscribers?
  553. - Exactly.
  554. This all changed.
  555. And Steve Kornacki was named
  556. one of the sexiest men alive guys.
  557. Come on, we're on top.
  558. Yeah, the dog agrees with that assessment.
  559. So we have talked a lot about COVID on this podcast.
  560. And I wonder what you think the longer term effects
  561. are gonna be for your field
  562. of that very data intensive experience
  563. of working with COVID data in a way that readers
  564. and users actually really responded to.
  565. - I mean, the tracker thing is here to stay.
  566. And I think the creating dashboards for the general public
  567. on major topics they care about is not new.
  568. I mean, it's something that's been around with weather
  569. and sports for a long time.
  570. But I think that it's a type of product
  571. that clearly our audience wants
  572. that is in my opinion, like capital J journalism
  573. in most circumstances, but requires a new assembly line
  574. to be constructed inside the metaphorical factory
  575. of your newsroom to create that.
  576. You know what I mean?
  577. It requires a little bit more
  578. of a product development mindset,
  579. a longer development cycle, a more iterative approach,
  580. you know what I mean?
  581. Than juggling what's on Sunday's front page
  582. and the sort of like day-to-day heartbeat of the newsroom.
  583. And so I think there's a way in which
  584. to produce more of those kinds of widgets,
  585. data journals and teams are gonna have to find a way
  586. to create a kind of app workflow
  587. that's able to support that.
  588. And I think that's becoming more
  589. like a product development team in some ways.
  590. - From a different perspective,
  591. I think it's gonna have a long-term effect in newsrooms
  592. because I think it really woke up a lot of people
  593. to the need and benefits to being able
  594. to analyze your own data
  595. and not rely on a government official
  596. to hand you some numbers on a press release.
  597. Especially in those early days
  598. where we kind of had to pull stuff together ourselves.
  599. But then even more long-term is data literacy,
  600. increasing understanding among reporters and editors
  601. for the need to understand
  602. all the weird little caveats behind data
  603. and the problems we have with denominators and whatnot.
  604. And I think there's gonna be,
  605. I'm expecting to see greater appetite
  606. for data training in my newsroom.
  607. - We talked a lot about how data journalism is affecting
  608. the bottom line for news organizations
  609. and how they're all seeing the value now.
  610. If you're a reader or a user,
  611. you don't necessarily realize
  612. you're looking at data journalism, I guess.
  613. But what difference does it make to them?
  614. Do you think it helps people understand
  615. the local areas better
  616. in ways that we haven't quite kind of grasped
  617. the long-term impact of yet?
  618. - Well, I think the way we present data journalism,
  619. especially in charts and graphs and things like that,
  620. allow people who maybe need a visual aid like that
  621. to see something rather than reading
  622. long sentences of text,
  623. that some readers just do better
  624. with absorbing information that way than others.
  625. And I think also the readers maybe
  626. who aren't naturally inclined to do that
  627. are getting better at it because there is so much of it.
  628. - We all have, oh, sorry, Ben.
  629. Were you going to say something?
  630. We can make a pause.
  631. Let's set this up, okay, perfect.
  632. I was about to say, let's begin there,
  633. we all have projects in our past as professionals
  634. that we are very fond of,
  635. even if they are not perhaps the best projects
  636. that we have ever done,
  637. but there are certain projects that we hold dear, right?
  638. So if you had to choose just one of your local
  639. data journalism projects in the past, right,
  640. what would that be and why?
  641. What was the story about?
  642. What do you do to gather the data or visualize
  643. or display the data, explain the data,
  644. and what type of impact that story had?
  645. - In 2018, we did a series called Denied Justice
  646. that looked at how the criminal justice system
  647. in Minnesota was handling sex assault investigations.
  648. A reporter came to me and said he was hearing
  649. that very few cases that get reported to police
  650. actually result in a conviction, but nobody tracks that.
  651. It's not a number that someone calculates.
  652. So we put in requests for thousands of police
  653. investigative reports and we built our own database,
  654. tracking key bits of information off these cases
  655. and then hunting down, was this referred to the prosecutor?
  656. Were charges filed?
  657. Was somebody convicted?
  658. And a mountain of work in order to get a few numbers.
  659. And what we found is like 8% of reported cases
  660. result in a conviction.
  661. And we ended up doing a series,
  662. it ended up being a nine part series,
  663. which we did not expect,
  664. because it ended up being so powerful.
  665. The response was overwhelming from readers and policymakers.
  666. And we had hundreds of victims come forward
  667. to share their stories because they read our first story
  668. and were like, oh my gosh, I'm not the only one
  669. who had this bad experience with the police department.
  670. I'm not the only one.
  671. This is so cathartic to hear this
  672. and they wanted to share their stories.
  673. And as a result, Minnesota's sexual assault laws
  674. have been overhauled quite substantially.
  675. Several police departments have added additional staff
  676. or like victim advocates embedded into their department,
  677. prosecutors embedded into the department
  678. to help improve the quality of the investigations.
  679. And it was one of the most,
  680. probably hands down the most rewarding
  681. piece of work I've ever done.
  682. At the end of it, we gathered all these victims
  683. at the state Capitol and took a picture
  684. of all of them together.
  685. And these women came out in the bitter cold in Minnesota
  686. in December to do this photo.
  687. And they came up to me and the other reporters
  688. and they were so thankful.
  689. And I've never had that experience before
  690. that the people just thanking us for the work.
  691. And that story is a really big combination
  692. of data being the spine holding up the key findings
  693. and then these powerful emotional stories from victims.
  694. And policy makers told us that the two things combined
  695. were so powerful that they could not ignore it.
  696. - Yeah, when the data comes together
  697. with real life in a way, you know what I mean?
  698. Or when the data leads you to the doorstep
  699. you didn't know you needed to go to, right?
  700. That is just this sort of like unforgettable feeling.
  701. You know what I mean?
  702. It's kind of like one of the deeper work hits.
  703. It's better than any hack, you know what I mean?
  704. And so like for me, like an example of that was
  705. we had a circumstance where there was like
  706. a little political fight in LA government
  707. over the fire department that was really pretty superficial
  708. and was just a couple of people
  709. trying to score political points.
  710. But it raised some questions about the efficacy
  711. of the 911 system in the city of Los Angeles.
  712. And so my colleague and I, Robert Lopez,
  713. decided to get a copy of the entire 911 calls database
  714. and analyze what was happening with the 911 system.
  715. And you know, there's a lot of ins and outs
  716. to how it all worked out.
  717. But the reality was is by doing our own
  718. independent analysis of the data,
  719. we were able to identify several structural flaws
  720. in how the 911 center operates,
  721. how they handle calls near the city border
  722. with neighboring fire departments,
  723. and then go out into the real world tracking real calls,
  724. knock on doors and find out what happened.
  725. And in that we discovered these kind of deeper problems
  726. with the system that nobody was talking about
  727. in city politics.
  728. And we were able to really point to where the system
  729. needed to improve to like make a difference.
  730. And then a change actually did happen, you know.
  731. And to me, that's sort of like that feeling of
  732. where the statistics lead you to the store,
  733. you listen to the numbers,
  734. and then you follow them out into the real world.
  735. And then you connect that with what's really happening there.
  736. And then you bring that back to your audience.
  737. That's better than any chart.
  738. That's better.
  739. You know, I love charts, you know what I mean?
  740. But like, that's better than any little dinky thing.
  741. That's the deeper like thing
  742. that numbers and statistics can lead you to,
  743. which is the sort of hidden truth,
  744. but the deeper story that isn't being told.
  745. And that's the hit, you know,
  746. I'm still looking for it on every story to get back to.
  747. I'm still chasing.
  748. - Those are both examples of stories where a human
  749. is probably not gonna come forward and tell us
  750. that there's a story there.
  751. Maybe they'd have an inkling,
  752. but probably nobody did the data analysis
  753. to know for sure that that's what it is.
  754. And it takes us to do these kinds of lifts
  755. to heavy lifts on data to find those stories.
  756. And, you know, as I said, interview the source
  757. to get at the meat of that story that needs to be told.
  758. - And then sometimes I think this may have happened
  759. in Mary Jo's case too.
  760. Your finding, your synthesis of what's really going on
  761. actually gives a name to and some meaning to something
  762. really negative that happened to somebody.
  763. And then they decide, then your sources
  764. and the people you're writing about
  765. find this deeper reward in the whole thing.
  766. And it's very rare, you know what I mean?
  767. But to me, that's as good as it gets.
  768. - That combination of like traditional reporting
  769. with the data and that sweet spot event have impact as well
  770. is just everything you can ask for, isn't it?
  771. So we are in February, this is our first part of the year.
  772. If you two were to make predictions,
  773. I know everybody does make predictions,
  774. but you two make predictions
  775. about what your big exercise is gonna be this year.
  776. - Let's let Mary Jo off the hook here.
  777. I don't want you to make her pop up too much.
  778. So since I'm working on this new project at Stanford
  779. called Big Local News,
  780. which is really an open source collaborative effort
  781. to try to be a data backbone for all the local newsrooms
  782. who maybe need help or some consulting
  783. or some technology to really pull off big data projects,
  784. we're looking to take it up a notch, Simon.
  785. I mean, I think this year, this team I'm on
  786. with Cheryl and Surdar and everybody up on the farm
  787. in Palo Alto, I'll say it here,
  788. we are going to build the biggest web scrapers
  789. in the history of data journalism.
  790. We're gonna, we have a project called Agenda Watch
  791. where we're gonna scrape hundreds of city council agendas
  792. across the country, put them into one database,
  793. and we're gonna try to do the same thing
  794. with county courthouses.
  795. And you know, those types of projects,
  796. scraping your local county courthouse or city council,
  797. I mean, every local newspaper in America
  798. has wanted to do it.
  799. Some of them have tried to do it.
  800. A very few have succeeded in doing it
  801. for like a short period of time,
  802. but I just don't think that there's been a lasting
  803. and sort of broad solution to that challenge
  804. because it's a lot of work, there's competing priorities,
  805. you know, da, da, da, da, da.
  806. And we're really looking to solve that or try to
  807. by being the sort of database library
  808. that attacks those big things.
  809. And so by December 31st,
  810. we'll be the biggest scraper you've ever seen,
  811. outside of Google, of course.
  812. - I love the Local News Project,
  813. I'm really glad to hear that.
  814. (upbeat music)
  815. Okay, so we are gonna finish off with a new section,
  816. which I'm gonna just call no dumb questions
  817. because these are quite dumb questions.
  818. And they're one word answers,
  819. but I'm gonna go first.
  820. And my one to Mary Jo is,
  821. what book or class had an impact on you
  822. as a professional?
  823. - The Investigative Reporter's Handbook.
  824. - Very good, how about you, Ben?
  825. - Oh, just one?
  826. - Yeah, just one, I'm afraid, sorry.
  827. - Oh, man.
  828. - Be concise, we are all journalists,
  829. we are supposed to be concise.
  830. - Gotta be concise, all right, I'm gonna do it.
  831. I will say the one that had an effect on me,
  832. I will say The Autobiography of Malcolm X
  833. by Alex Haley, guys.
  834. How about that?
  835. It's like a young teenage me,
  836. you know what I mean, moving from the country to Chicago,
  837. getting a broader view of the world.
  838. It was nonfiction journalism
  839. that helped me like see what was up,
  840. you know, and its book is an alternative version
  841. of American history that I hadn't heard before.
  842. It's also a religious conversion story,
  843. not unlike one I've been taught in church,
  844. you know, as a kid who went to Sunday school.
  845. And it shows the power of stories
  846. that go against the grain
  847. that aren't otherwise being told.
  848. And it's a great piece of literature.
  849. So if you haven't read it, you know,
  850. you should check it out.
  851. I think sadly in some ways, you know,
  852. we've regressed since then.
  853. You know, that was a time when Alex Haley
  854. could have the most popular show on television
  855. and write a really provocative book.
  856. And as a reminder of the power,
  857. and I would say virtue of the mainstream media,
  858. when it's properly focused, you know.
  859. - Those were the times.
  860. Okay, next quick question for both of you, actually.
  861. So if you could keep just one of the tools
  862. that you commonly use in your daily job,
  863. what would that tool be?
  864. - R.
  865. - I like that answer.
  866. What about you, Ben?
  867. - Oh, a spreadsheet, guys.
  868. It all begins with a spreadsheet.
  869. You can't play baseball without a bat
  870. and you can't crush data without a spreadsheet.
  871. And--
  872. - Oh, you can, you can, R can do everything.
  873. There's so many things that R can do that Excel can't.
  874. - These R people, they really love R.
  875. Have you guys noticed this?
  876. Like, I think--
  877. - I have actually noticed that.
  878. - Yeah, I think that, you know,
  879. not unlike, that it's like R conversion stories.
  880. It has an evangelical fervor and flavor,
  881. not unlike the Nation of Islam.
  882. - You are only saying that because you're a Python person.
  883. That's the only reason why you're saying that.
  884. I thought you were gonna say Python for sure.
  885. - No, I think it all comes down to the fundamentals.
  886. You know, you can't dribble.
  887. You can't play basketball if you can't dribble.
  888. Can't play baseball without a bat.
  889. I'll give you the religious flavor,
  890. is I'll pick Libre Office's Calc,
  891. the open source free alternative spreadsheet,
  892. which is just as good as the ones from those companies
  893. you maybe have heard of.
  894. - Okay, if you rephrase the question to
  895. what should any journalist have a tool,
  896. I would say Excel.
  897. But what should I, as a professional data journalist,
  898. doing this full-time, crunching numbers like non-stopped,
  899. then it's R.
  900. - All right, okay.
  901. Next one, what is your favorite hack
  902. that you are too embarrassed to admit?
  903. I can tell you my one is,
  904. and I still do this occasionally,
  905. is the way you can use the finder and place
  906. in Microsoft Word as a shortcut to getting stuff done
  907. if you're in a hurry.
  908. I still use it.
  909. I'm slightly embarrassed about it.
  910. So there you go, that's mine.
  911. Ben, do you go first?
  912. - Well, I'm not easily embarrassed,
  913. which maybe is a bug, not a feature in my personality.
  914. But I mean, a hack,
  915. saying no to other people in the newsroom.
  916. You know what I mean?
  917. I think that the data journalists,
  918. especially in smaller newsrooms,
  919. can sometimes get consumed by helping things
  920. just go the way they've always been going
  921. or sort of chipping in around the edges.
  922. And I think that being a helper
  923. and assisting that is noble work and often really valuable.
  924. But I see it pulled back a lot of other data journalists
  925. from really reaching their potential from taking risks
  926. and from doing the more ambitious, powerful work
  927. that our skills are capable of.
  928. So never be afraid to say no
  929. and never be afraid to bet on yourself
  930. and to take a chance.
  931. - I guess I don't know if I'd call this a hack,
  932. but I am probably most embarrassed to admit
  933. that I have data import scripts
  934. that I have not automated, but I probably should.
  935. And I just still keep running them.
  936. It's not exactly by hand, but they're not automated.
  937. I feel like I should be moving into the 21st century,
  938. but, oh well.
  939. - Well, I still run out of patience
  940. whenever I need to style graphs or charts in R
  941. and I export them as PDFs and I style them in Illustrator
  942. just because I find it faster.
  943. So that would be my personal hack.
  944. And I'm not embarrassed at all in admitting it.
  945. So we all have a skeletons in our closet, I guess.
  946. - I had trouble making a graphic in R yesterday
  947. and I was like, ugh, okay, I'm just gonna put it in Excel
  948. and I will make it in Excel.
  949. - I do the same thing with Illustrator.
  950. It's like R is not working, fine.
  951. Let me go to the Illustrator's graphing tool,
  952. which is sort of like the clunkiest graphing tool ever.
  953. So anyway, so let's go to the next one.
  954. What would you have done
  955. if you hadn't gone into journalism?
  956. - Oh, MG, I have no idea.
  957. I've wanted to be a journalist since I was 16, no clue.
  958. - Well, I was not in a very motivated place
  959. before I found journalism as we covered already, Alberto.
  960. But one thing I did like to do at that time in my young life
  961. was read pretentious books and be like a pretentious student.
  962. And so I think at that time, if you had asked me,
  963. I probably would have told you
  964. I would have become a high school English teacher
  965. or something, you know what I mean?
  966. And I don't think that that's a crazy outcome for me.
  967. And maybe I would have been happy doing that.
  968. It's hard to say.
  969. I probably would have stayed in Chicago
  970. and now that I think about it, never met my wife,
  971. which is making me sad to even contemplate.
  972. Like, you know, that's probably it.
  973. - So glad you didn't do that, Ben.
  974. (laughing)
  975. - Oh my God.
  976. My monologue to like high school students
  977. about of mice and men, it would be just terrible.
  978. - I would hope I need to hear that.
  979. Okay, last one, pie charts or tree maps.
  980. And I'm going to go to Mary Jo for this one.
  981. - Tree maps.
  982. - Like anything with visualizations, it depends.
  983. You know what I mean?
  984. Like, so like, I would say like,
  985. well, at first I would say never pie charts
  986. in the circular form.
  987. However, if you have like a single,
  988. like a categorical thing,
  989. I think the like underrated pie chart alternative
  990. is a single stacked bar.
  991. You know what I mean?
  992. Just like one rectangle sliced up.
  993. You can do the nice little labels on the top and bottom.
  994. And I guess it's technically a stacked bar chart,
  995. but it's just a single bar.
  996. It's also a pie chart.
  997. I think those are great.
  998. I think those really work and you can stack them.
  999. It doesn't like Pew do that on their quizzes.
  1000. I think that's good.
  1001. Cause like, cause maybe that is a tree map.
  1002. I don't know.
  1003. You guys tell me in some way it is a tree map, I guess,
  1004. but like to me the tree map works
  1005. when there's a lot of slices,
  1006. like there's like 50 slices
  1007. and like one or two of them are huge, you know?
  1008. And you want to be like, these ones are big
  1009. and these ones are small.
  1010. But like, if you only have like five things, you know,
  1011. you don't need that vertical part of the tree map.
  1012. If it's just like a fixed wide height, it works fine.
  1013. Sorry, I just did that.
  1014. But that's really how I feel about it.
  1015. - It's a great, it's a great answer.
  1016. I mean, tree maps were created to show a nested hierarchies,
  1017. right?
  1018. So you have a total
  1019. and then you subdivide that total into large divisions,
  1020. for example, population of the world.
  1021. And then you start dividing them into continents
  1022. and each one of the continent pieces
  1023. gets further subdivided into the country.
  1024. So that's what tree maps are for.
  1025. So it's an excellent answer.
  1026. I would advocate for the pie chart though.
  1027. So just to be the contrarian,
  1028. I would vote for pie charts if I could.
  1029. Simon, can I do that?
  1030. - You don't get advocate for pie charts.
  1031. - I did not expect that, Alberto.
  1032. - I'm getting old.
  1033. Getting older, you know, on a simple pie chart.
  1034. I tweeted the other day, if a graphic works, it's good.
  1035. And pie charts unfortunately are fortunate, often work.
  1036. It's like if there are two, three subdivisions,
  1037. four subdivisions perhaps, readers like them,
  1038. readers understand them, readers read them,
  1039. they bring attention to the data.
  1040. So, you know, I'm getting older, more flexible,
  1041. more pragmatic in some sense.
  1042. So go for the pie chart.
  1043. - Even the three-dimensional pie chart?
  1044. - No, three-dimensional.
  1045. That's going to, maybe when I'm 80.
  1046. - That's the limit so far.
  1047. - Yeah, that's, no, maybe when I'm about to retire,
  1048. maybe I will accept three pie charts.
  1049. - Wow, now that Alberto has been switched with somebody else,
  1050. I think that's a good point to end.
  1051. Ben Welsh, MaryJo Webster, thank you so much.
  1052. - Thank you.
  1053. - Thank you for having us.
  1054. (upbeat music)
  1055. (upbeat music)
  1056. (upbeat music)

Downloads

Recording audio · Timestamped transcript