Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Episode #89: Priya Krishnakumar
172 Plays8 years ago
In this week’s episode of the PolicyViz Podcast, I’m joined by Priya Krishnakumar, a visual/data journalist at the LA Times. I’m a big fan of Priya’s work and her colleagues at the LA Times. For me in particular, I think...
The post Episode #89: Priya Krishnakumar appeared first on PolicyViz.
Recommended
Transcript
Introduction of Priya Krishna Kumar
00:00:11
Speaker
Welcome back to the Policy Viz podcast. I'm your host, John Schwabisch. So here we are. It's sort of almost feels like the middle of summer here in DC. Hopefully everybody else is cooling a little bit. But I'm excited this week to have on the show Priya Krishna Kumar. See Priya, I was able to do it. It is phonetic. Well, actually, what is your formal title at the LA Times?
00:00:34
Speaker
My formal title is Graphics and Data Journalists.
Priya's Background and Role
00:00:38
Speaker
Graphics and Data Journalists. Okay, well, interesting. Well, welcome to the show. Thanks for coming on. Thank you for having me. So I think we have a lot of stuff to talk about, both sort of specific projects and maybe some personal philosophies, shall we say? But maybe, why don't we start with having you talk a little bit about your background and how you ended up at the LA Times?
00:00:59
Speaker
Yeah. Okay. So, um, I went to college at Northwestern, um, in Chicago and while I was there, I got pretty involved with their night lab, which is, um, it's like a journalism program, uh, funded by Mozilla that kind of lets people in the journalism school and not in the journalism school experiment with like the intersections of tech and journalism. So I got kind of involved with that and that has just kind of led me to basically all my opportunities since then. Um, I,
00:01:29
Speaker
Interviewed for my current job while I was at night car in college, which I went to because of the night lab and I have been here since I graduated college for
00:01:40
Speaker
Almost three years now. So yeah, it's been awesome That's
Photography and Visualization Techniques
00:01:43
Speaker
great. And like the your website has it's basically like infinite scroll almost of awesome projects from times I think there are two big topics in my head that I want to get your thoughts on so the first topic is You seem to use a lot of photography
00:02:01
Speaker
in your work, sort of before and afters, small multiples, sliders, and that sort of thing. So I want to get your take on how you use photography, how it relates to data journalism, or even maybe we need to define data journalism first. So that's the first topic. And then the second topic is this balance or relationship, balance I think is the right word, between interactive visualizations and static, and how you think about using those in your work.
00:02:29
Speaker
So maybe we can start with the photography because I feel like one of the things I really like about your work is how much photography you use in these sort of real data driven ways. So can you talk a little bit about how you think about using photography in the way that you do at the newspaper? Yeah, I mean, I personally think so. I'm a terrible photographer, but I love photography and we have such an incredible photo staff here.
00:02:57
Speaker
that any chance that I can to like collaborate with them I will like jump at because and I also think that a lot of our stories are like data stories can really be enhanced by using photography especially like California where so much of what we do is focused on like the environment and the drought that we have this like huge wealth of information in these photos like why would we not use that.
00:03:21
Speaker
And I'm trying to think of some other examples too, besides just the environment. We did a bunch of stories over the summer. I did a big infographic about unrest at Trump rallies and we had photographers who were able to capture that so much more than like a chart of, you know, a chart can really convey so much emotion. And I can tell you how many people were arrested and where they were arrested. But if you see a photo combined with that, then you actually feel something. And I think if you can make people feel something
00:03:52
Speaker
while they're also getting some data, that's good too, I think.
Engagement with Photo-based Visualizations
00:03:56
Speaker
Do you feel that like for the sliders and the interactivity that you've used a few times, you know, do you feel like people might engage with those more than an interactive chart or just maybe differently than a data visualization when it's a photo interactive visualization? I definitely do think that people approach it differently. I think it's just
00:04:20
Speaker
It's different than what you'd normally expect to see, so you might be a little bit more compelled with it, especially if it's combined with a chart. We did a project a couple weeks ago that compared Bush and Obama and Trump's first 100 days, and it was a quiz. And the first quiz was the two photos of Obama's and Trump's inauguration side by side. And that in itself is data. You can see the numbers in the crowd, but it's still a photo.
00:04:50
Speaker
I don't know, I just think that people come away with something a little bit different than they would with like a chart or...
00:04:56
Speaker
Yeah, I think the engagement is different because you see actual people as opposed to data points which means different things. What are some of the challenges you found with working with photography either in terms of selecting a photo or their quality relative to using data where you sort of have in some ways more freedom to just make a chart and your choices are aesthetic and what chart type to
Accuracy in Data Visualization
00:05:23
Speaker
use, right?
00:05:23
Speaker
I mean, I think you still have to approach it with the same level of accuracy and detail that you would with a chart. Like for the photos of the Trump and Bush inaugurations, I made sure that I overlaid the photos on top of each other before I put them side by side to make sure that they were in the exact same frame and that, you know, everything matched perfectly because you can still mislead, just in the way you can mislead people with data or chart, you can mislead them with a photo too.
00:05:52
Speaker
If you're incorporating a photo into a data journalism project, I just think you need to be really cognizant of that while you do it. Yeah, absolutely. The same goes with sliders too. If you do a slider with two images that don't exactly match, I think it's kind of pointless.
00:06:07
Speaker
Yeah, that's really interesting. And do you work with the photo journalists there to say, Oh, we have this concept of a story that we want to do a before and after the drought, for example, you did this piece of the drought. Do you say, you know, we kind of want to get this picture so we can line it up? Or is it just happenstance that, you know, you get photos and they're like, Oh, I remember this other thing and we can line these up now. Um,
00:06:31
Speaker
Well, for the drought ones, we found them on our own. I think we use satellite imagery for that, and that's something we can do on our own. But a lot of the times, if I'm looking for a photo, I will just shoot our photo department head an email, and he'll get back to me within two minutes. They're always eager to help us out and work with us.
00:06:49
Speaker
It helps all of us, I think, when we can work together like that. Right, absolutely.
Static vs. Interactive Content
00:06:54
Speaker
Cool, OK. Well, the other big question or topic, I think, in the field of journalism and data journalism is the balance between making something interactive versus making something static. You had a piece, I think it was last October, about how California switched from red to blue.
00:07:18
Speaker
And it was a whole bunch of sort of small multiples of California, and they were all static. And then only at the very end of the article was a sort of interactive data table where you could go in and put in your zip code or your city. So I'm curious about your philosophy, your thoughts on the balance between making something interactive versus making something static. I think a lot of the times, I think lately I've been trying to hew more towards static if I
00:07:43
Speaker
And I think the reason is a lot of the times doing like tool tips or interactivity can hide a lot of the pertinent information from a reader. And with something like that California small multiples thing, like it's a pretty basic, you know, red, blue or independent thing. So you kind of get it as soon as you see it. And then if you want that level of additional detail,
00:08:08
Speaker
I think it kind of just depends because with that one, it's like if you do a tooltip, do you hover over one county and then you see all the tooltips for like the rest of the counties? And I thought that would be like a little too overwhelming. So then at the bottom, there's like a table where you select the county and then you can just see it by year. And I thought that would maybe be more effective. I think also a lot of the times is it's if you rely too much on
00:08:32
Speaker
tool tips and things like that, you get into kind of a quandary when you move it to a phone. And the reality is that I think that page was viewed, I think more than half the people who read it looked at it on their phone. And I feel pretty strongly that it's very difficult to do a good tool tip or like a lot of complicated interactivity on a phone. And so a lot of the times, and I think that like a more static experience translates pretty well to a desktop too, you know?
00:09:00
Speaker
Are you thinking mobile first for the most part these days? I try to. Sometimes it gets pretty tough because we all have these like huge monitors and whatever to just like design for a huge desktop. But that's just not how people read our stuff. That's how I read my stuff and how I read the stuff of my peers. But I'm not you know ideally you don't
00:09:21
Speaker
create data visualization for your peers, you know, it's like our readers, right?
Effective Visual Storytelling
00:09:27
Speaker
So when you're thinking of your reader, when it comes to interactivity, what is the picture you have in your head, because I often feel like people make interactive visualizations, and it's really just allowing a user to go to the second level of diving in and getting all the details
00:09:43
Speaker
But I don't think that's oftentimes what you know, regular readers, they just want to get the they want to get the overarching story and they're not going to dive in. So when you're thinking about your reader, what's the avatar, I guess you have in your head. My assumption is that people are usually in a hurry. It's hard to like really capture a reader's attention and expect them to sit down. It's a big ask of anyone to sit down to one of your projects and spend like 10 or 15 minutes
00:10:11
Speaker
reading it and really getting into the leads. I really doubt that that's happening. And if it is, that's awesome, but you really can't expect that. So I want a reader to come to a page I've done and just get it. If they spend one or two minutes on it, I would hope that they've kind of gotten what I wanted them to get out of the story. And then if they want to spend more time with it, I want that information to be available to them, but it doesn't necessarily have to be
00:10:38
Speaker
the first thing they see, I guess. Yeah. So it's a cost benefit analysis in some way, right? Because there's some there's some subset of your readers who are going to want to dive in. But it's going to take you I would guess maybe not you but for most people I would guess like to create the interactive piece. I don't know, maybe I'm wrong about this. But I would guess to create the interactive piece is going to take a bit more time and investment to create the static piece.
00:11:03
Speaker
So as you're thinking about that different reader who's trying to get the data and engage in that way, are you thinking, well, I'll just find a different way to just give them the data as opposed to making this huge interactive visualization and just here's the GitHub page or here's a table with the data? Yeah, I mean, I don't think that the solution is to not provide that data, right? I think it's just about
00:11:29
Speaker
maybe giving away that kind of lightens the load on everybody, right? Like I made that graphic during election season and it's like I could spend like another, you know, day or two making like a perfect tool tip that will also look great on a phone or I could like give everybody like a good top level overview, a table at the bottom and like move on to the next thing. You know what I mean? And like the information is still there, but then
00:11:53
Speaker
I can do more. The readers hopefully still got what they wanted to get out of it. Yeah. So when you are thinking about documenting your sources plus the methodology and providing that background to readers, how are you or how is LA Times in general approaching
Data Sharing and Public Service
00:12:12
Speaker
that? So for example,
00:12:13
Speaker
FiveThirtyEight has a GitHub page, you know, and they're doing a lot of, you know, explicit modeling and things. But how are you guys approaching your sort of openness when it comes to the analysis side of data and the code behind the data that you've used to scrape from a website and then build this other thing out of it? We have like internal discussions about what we want to be like open source and what we want not to. So for example, two of my colleagues, Joe Fox and John Joyce. So the
00:12:43
Speaker
we all kind of had this big department-wide effort to do this big map of the election results of every precinct in California, which is a huge undertaking. And we all ended up getting involved. But then at the end, John and Joe were the ones who went back and made sure we had all the finalized data and got all of that. And it's like every county in California and every precinct within every county, that's a huge amount of data. And with that, I think since we have all of it,
00:13:11
Speaker
It became like, you know, it's a public service to make all of this available and public. And here's how we parsed it. And like with that, and that's helpful, not just for us as a department, but for all the journalists in our newsroom who can like now take that data and use it to find stories, but like journalists all across California or like whoever, you know? So I think when it's beneficial to a larger purpose, I'd say is when we do it.
00:13:38
Speaker
When you talk to other folks in the newsroom, so you're a graphics and data journalist. So what's the relationship like or what's the, how do you work with the journalists who are maybe don't have that data in front of their title name? Like how's the interaction and how's the workflow with the rest of the newsroom?
Collaboration in Newsrooms
00:13:56
Speaker
I think it's really collaborative. I think it's become more collaborative in the past few years to even that I've seen. And that's been really like awesome and helpful because
00:14:07
Speaker
For example, if I look at transportation data, I can look at it for days and maybe get some meaning out of it, but I still won't come at it with the same context and nuance that our transportation beat writers would. And I think especially in our newsroom, we have so many reporters who are so well-versed in so many different niche aspects of the city and state that it's beneficial both to me and to everybody else if we can just
00:14:37
Speaker
Collaborate together and I think most of our reporters in the newsroom know that like a Data source or like a good data source can help like enhance their story and then if we can work together You know, we can create something that's
00:14:51
Speaker
beneficial to like everybody and our reader. Right. So I think it's been good. If science remains any sense. I don't know. I think we have a good relationship with people in our newsroom. I mean, I think the transportation data sets are a great example. So you have a transportation data set, maybe you talk to the transportation journalist or writer. How much are you trying to educate them on
00:15:12
Speaker
the sort of what it takes to work well with data and to and what you can create with the visualization but i'm thinking sort of first step of like. Yeah i can give you the mean of this variable but like there's these outliers and and so i wonder like how much education is going on.
00:15:29
Speaker
you know, both ways, obviously, but going on to educate the journalists who may not have a strong statistics or math background on, you know, ins and outs of data, or maybe, maybe the journalists of the LA Times are all really good at math. I don't really know. But I'm just curious about how that, you know, how the different skill sets are sort of matching up and the information sharing and knowledge sharing is working at the newsroom.
00:15:52
Speaker
Um, like sometimes what will happen is like a reporter will get a dataset and not necessarily know like what to do with it or like how best to parse it. And so they'll bring it. So we have a graphics desk and a data desk. Um, and we all, we all work together, but we have like people who are real statisticians on the data desk, which is not me. So a lot of the times they'll bring it to them and then like, I'll come in and then like work with one of them. And then we'll like all work together to figure out what's like the best way, the best thing to do with the data that we have. So it's like really collaborative. And I think.
00:16:22
Speaker
If I have a big gaping hole in my knowledge, somebody else will have that, be able to help me out and help the reporter out. So I think it's been, it's been pretty good that way.
00:16:32
Speaker
Yeah, it sounds like what you have there is the team that you need to sort of have all the key skill sets you need to tell these stories. Is there a spectrum of data driven stories? And does the workflow differ between those certain types? So where you might have a real data driven thing that you got some cool precinct level data, for example, and you're basing out of there versus
00:16:58
Speaker
or not versus, but on the other side where you have a journalist who, you know, went out and interviewed some people and want some data to buttress or support the story that they're telling. And is there a real fundamental difference in how the newsroom approaches those stories? Or is it all just collaborative and it all just sort of works out in the end? Um, I think it all kind of works out in the end. Like, yeah, I don't know. Like, I think
00:17:23
Speaker
someone's like reporting out of Houston and they want census data for Houston, like one of us can pull that really quickly and help them out. But at the same time, you know, we're usually working on like multiple things at once. So I can like do that quickly for somebody while also working on a longer term data project. So usually things kind of work out. Yeah, yeah, yeah.
Working on Wildfire Data
00:17:40
Speaker
Do you have any longer term data projects you're working on right now that you want to sort of talk about either quietly without giving away the headline story, talk about maybe some of the methods that you're and challenges that you're having?
00:17:52
Speaker
Um, yeah, I think something that like, we have been working on a little bit is how to deal with the massive amounts of wildfire data that we have available to us. Um, and I think that's something that, you know, other newsrooms in California have, they've done a really good job of, um, I think like KPCC and other, you know, other stations have, have done a really good job, but we have all these
00:18:19
Speaker
fires in California and we have all these places like NCWeb and wherever who collect all the state about where the fire is and what's going on, what the burn marks are like. And I think what ideally we'd like to do is take all of that and figure out how to use it in a helpful and interesting way that we can sustain hopefully in the years to come because there are always a bajillion fires every summer.
00:18:49
Speaker
So I think that's like the next big thing that.
00:18:53
Speaker
I think we would like to tackle and I would like to tackle. So wildfire seems like a natural one. That's going to recur
Project Sustainability vs. Reusability
00:19:00
Speaker
every year. But are you generally trying to build as you're building an interactive, you're building a table, are you generally thinking about how you're going to be able to maintain that, not to maintain it or extend it or use it in another thing and have the same sort of approach? Or is it just most of the things you're doing are just a one off? I want to make this map. I'm going to go on to the next thing.
00:19:21
Speaker
I don't know that this is the best way to do it, but whenever I do a project, I usually approach it like it's a one-off, which is maybe not the best thing, because a lot of the times I will do a project and be like, oh, I could reuse this later. And then I have to go back and fix it so I can make it something I can reuse. Whereas maybe if I had approached it from the beginning as something I could reuse, it would make my life easier later on. But I think a lot of the times in the pressure of a deadline, you're just like, blah, let's get this out.
00:19:49
Speaker
But I think a lot of the stuff that we do that are one-offs end up becoming tools or reusable projects later on. Interesting. So then it's just like a question of, I guess, approaching the story. Like, do you approach a story with like a
00:20:02
Speaker
template and mindset? Or do you just kind of go back and fix it later? I don't know. I'm still trying to figure that out. I don't know if you have thoughts on it. Well, no, I think that the templates are great, but they also in some ways lock you in to sort of like, here's the path that I'm going to take, you know, a map is a good example, you have a map of California counties. And yeah, you could reuse the code. But then does that automatically lock you into that same kind of map? Whereas if you started from scratch, maybe you would end up
00:20:31
Speaker
in a different map type or maybe not using a map at all. So I agree. I go back and forth on whether I should be trying to template everything that I'm working on or whether it's everything I'm going to start fresh. So I'm starting fresh. Obviously, reusable code is helpful because you're not starting on the code from scratch. But yeah, I don't have a great answer either. Right.
Empowering Journalists with Tools
00:20:58
Speaker
I think database should be
00:20:59
Speaker
interesting and like fresh and I don't know like our readers have seen yeah like if a reader has been with us for like five years they or whatever they've seen the same map probably like a bunch of times the same chart a bunch of times and like how do we keep them engaged I don't know that using like the same template a bunch of different times is like in their best interest or ours really you know right
00:21:22
Speaker
Yeah, I don't know the answer to the question, but you know, and I think there's different types of content producers who it helps and hurts. So I think of a case where you might want to empower other people to be able to make an interactive map, for example. Right. And in that case, you might want to give them the template because they just, you know, they're not the expert, but they would just want to drop their new data in and
00:21:43
Speaker
go on to the next thing. Whereas if you are initiating the content creation, you might want to have control from the very beginning and be able to think about different ways and start from there. I don't know. Yeah. If we find ourselves doing the same thing over and over and over again, I think we try and make it so that we can hand some of that off. We have a tool that helps reporters make simple charts and things like that and simple maps.
00:22:11
Speaker
And that's kind of helped us because we can still help them out with that, but it also frees us up to try some of the new stuff that we want to try. Right. Maybe my last question. So do you think that's the new wave of things in the newsroom is having these custom internal tools? It seems like everybody now has their own custom internal tool. And I guess the secondary question to that is,
00:22:36
Speaker
Are newsrooms talking to each other about their how they built their internal tools and what's in it and how they can improve it and that sort of thing? Or is it on the LA Times, you're the New York Times, you're the Washington Post, we're all like, yeah, we're all friends, but we're not that friendly. Um, I think we should be talking to each other. And I think we are on some things like we use a fork of courts as chart builder tool. Right. And that's how our reporters make their like simple charts. And I think that's been like,
00:23:02
Speaker
really transformative for us because it like empowers reporters to find data and use it and see like, oh yeah, I can do this, you know? And like, I think there are other newsrooms that use it as well. And I mean, I don't see why we shouldn't be doing that because it helps all
Conclusion and Thanks
00:23:19
Speaker
of us. Like if I'm like a graphics journalist somewhere else, like I want to be able to, you know, also have the time to like experiment and try weird stuff and like, you know,
00:23:28
Speaker
And if we can all make that easier on each other, that sounds very like kumbaya, but I think it's important. Well, here's for the kumbaya moment of data journalism. So, um, yeah, I'm down with that. Um, cool. Priya, uh, thanks so much for coming on the show. This has been a lot of fun. Yeah. Thank you for having me. Um, and thanks to everyone for tuning into this week's episode. If you have comments or questions, please do let me know. So until next time, this has been the policy of this podcast. Thanks so much for listening.