Resources
Mar 20, 2025 |
57 min
|
617 views

Data Science at the Census Bureau | Jessica Klein | Data Science Hangout

To join future data science hangouts, add it to your calendar here: https://pos.it/dsh - All are welcome! We'd love to see you!

We were recently joined by Jessica Klein, Data Scientist at the Census Bureau, to chat about open-source adoption at the Census Bureau, the growth of R and Python communities within the agency, training opportunities and internal support systems, and working with complex census datasets.

In this Hangout, Jessica talked about the agency's move towards the Cloud and its efforts to reduce reliance on proprietary software like SAS and ArcGIS. She shared insights into the internal support system, mentioning training programs, workshops, and resources for using the Census API. She also discussed the challenges and strategies involved in transitioning from SAS to R & Python, including the importance of double coding for validation with mathematical statisticians and the increasing use of version control with Git GitLab. The development of internal technical support groups for R and Python was a key element in fostering knowledge sharing and breaking down silos within the agency.

Resources mentioned in the video and zoom chat:
πŸ”— Jessica Klein's LinkedIn β†’ https://www.linkedin.com/in/jessica-klein-8a5a35196/
πŸ”— tidy census package β†’ https://walker-data.com/tidycensus/
πŸ”— tigris package β†’ https://github.com/walkerke/tigris
πŸ”— Fast Link package for probabilistic linking β†’ https://github.com/kosukeimai/fastLink
πŸ”— StatQuest YouTube Channel β†’ https://www.youtube.com/channel/UCtYLUTtgS3k1Fg4y5tAhLbw
πŸ”— 3Blue1Brown YouTube Channel β†’ https://www.youtube.com/c/3blue1brown
πŸ”— {sf} package for R spatial analysis β†’ https://r-spatial.github.io/sf/
πŸ”— Resources on spatial analysis in R β†’ https://rspatial.org/ and https://rspatialdata.github.io/index.html and https://asdar-book.org/
πŸ”— Apache Superset (open-source visualization tool) β†’ https://superset.apache.org/
πŸ”— Coding it Forward Civic Digital Fellow program β†’ https://blog.codingitforward.com/tagged/civic-digital-fellowship

Resources shared directly by Jessica in the chat:
πŸ”— Analyzing US Census Data: Methods, Maps, and Models in R by Kyle Walker: https://walker-data.com/census-r/index.html
πŸ”— Webinar series on using tidycensus for analyzing and mapping Census data: https://ssdan.net/events/the-2025-ssdan-webinar-series-2023-acs-data-with-r-mapping-tools-and-the-2020-census/
πŸ”— A Guide to Working with US Census Data in R: https://rconsortium.github.io/censusguide/
πŸ”— Using the API requires a getting an API Key, which you can get for free here: https://api.census.gov/data/key_signup.html
πŸ”— See here for the Census Data API User Guide: https://www.census.gov/content/dam/Census/data/developers/api-user-guide/api-guide.pdf
πŸ”— Explore the ABCs of the Census API with Census Bureau staff: https://www.youtube.com/watch?v=JbThy7GUg3k
πŸ”— See here for a tutorial on Census API Basics: Simple Steps to Better Data Access: https://www.youtube.com/watch?v=d-FJ2IyVfdk
πŸ”— See here for a tutorial on Using the API to Get All Results for an ACS Table: https://www.census.gov/library/video/2020/using-api-all-results-for-acs-table.html
πŸ”— See here for a tutorial on Using the API to Get Results for Multiple Estimates: https://www.census.gov/library/video/2020/using-api-results-for-multiple-estimates.html
πŸ”— Explore ACS Data Stories–Stats in Action! to see how the public is using ACS data: https://www.census.gov/programs-surveys/acs/about/acs-data-stories.html
πŸ”— Visit the Slack show-and-tell channel and show how you use Census data: https://app.slack.com/client/T6AL55003/CRY4R9R7S

If you didn’t join live, one great discussion you missed from the zoom chat was about the enthusiastic reception and usefulness of the tidycensus package for accessing and working with census data in R, with many attendees expressing their appreciation and sharing resources related to it. Let us know below if you’d like to hear more about it!

β–Ί Subscribe to Our Channel Here: https://bit.ly/2TzgcOu
Follow Us Here:
Website: https://www.posit.co
Hangout: https://pos.it/dsh
LinkedIn: https://www.linkedin.com/company/posit-software
Bluesky: https://bsky.app/profile/posit.co

Thanks for hanging out with us! πŸ’›

image: thumbnail.jpg

Transcript#

This transcript was generated automatically and may contain errors.

Alrighty, welcome back everybody to the Data Science Hangout. If we have not met, my name is Libby. I'm a data science community manager here at Posit and I'm also a really passionate data science educator. I have a background in business statistics and experience teaching Python and R. I am based in San Antonio, Texas where right now it's a lovely 65 degrees outside.

I would love to know where you are in the chat. If you aren't familiar with Posit, Posit builds enterprise solutions and open source tools for people who do data science with R and Python. We are also the company formerly called RStudio, so if you know RStudio, you know us. I am joined by some wonderful people in the wings here who are going to be helping me out like Isabella and Curtis. Rachel, who started the Hangout many years ago, just popped up in here as well, so let me have Rachel hop in really quickly and introduce herself.

Sorry if I'm out of breath. I was running from an event in Boston. Hi everybody, so nice to see you. I'm Rachel. I lead customer marketing at Posit and I'll be hanging out here behind the scenes too.

Thanks Libby. Awesome. I was going to try to add in a fun fact about myself today. I listened to a podcast where every time the hosts give a fun fact and I couldn't think of one. What's a fun fact? I invite everybody to put a fun fact in their introduction in the chat this morning.

Well, the Hangout is our open space to hear about what's going on in the world of data across a bunch of different industries. It's one of the things I love about the Hangout is that we're all coming from different places, yet we share a lot of the same experiences and that can be such an enriching thing. So we chat about data science practices, data science leadership, we connect with other people who share our experiences, and we get together every Thursday, same time, same place, here on Zoom. So if you are watching this on YouTube later and you want to join us live in the future, I highly recommend it. The chat is amazing and you don't want to miss it. There's going to be details below in the description box so you can add it to your calendar. Just make sure if you add it to your calendar it adds for Thursdays at 12 p.m U.S. Eastern time.

Double check it once it adds to your calendar. I really want to thank everybody who has made this place the friendly and welcoming space that it really is today and we're all super dedicated to keeping it that way. So if you have feedback about your experience here today, whether that's good or bad, we really want to hear from you. There's going to be a survey that pops up very nicely in Zoom in your browser once you leave this meeting. So please fill that out to tell us how we're doing or share your thoughts. You could also find Rachel and I on LinkedIn and on Blue Sky. I think we're the only Rachel Dempsey and Libby Herron on both of those platforms.

All right, at the Hangout we love hearing from you. It doesn't matter what your years of experience are, your titles, your industry, the languages that you use or do not use, and we really encourage you to connect with each other in the chat. So introduce yourself, say hi, give a fun fact, include your role, where you are, something you do for fun, and please feel free to share your website, your LinkedIn, any resources. The chat is for us to hang out with each other.

Last time we did questions all on Slido instead of the chat and it was so good. It was fantastic. We got to add questions to Slido. You can put your name in there on Slido as well. So look for the little spot that asks you for your name. You could also leave it anonymous. That's totally cool. I encourage you to leave your name. You guys ask great questions.

And that allows us to upvote and sort of see what questions we want to add. And there's also a little reply button on those Slido questions, which is kind of cool. If you have a question that's for the group, feel free to continue asking that in the chat. And if it is for our guest today, put it in Slido. If you can't ask the question yourself live, if you weren't able for me to call you in live to ask, just put an asterisk somewhere in your question and I'll ask it for you.

All right. I want to make sure that Slido link has gotten to the chat. Do I have confirmation that that happened? There we go. Okay. Awesome. Just made it in 10 seconds. Sweet. All right. Rachel, with the Slido link in the chat, we will try to post that in the chat regularly so that everybody can see it. And, oh, I guess I should go see it. Hold on. Let me go get to the Slido.

Or else I'm not going to know what questions you're asking either.

All righty. I am so excited to be joined by our featured leader today, Jessica Klein, a data scientist at the Census Bureau. Jessica, I would love it if you could introduce yourself, tell us a little bit about your role and what you do outside of work for fun. Absolutely. So, hello, everyone. My name is Jessica Klein and I am, as Libby mentioned, a data scientist at the Census Bureau, but I didn't start out that way. So, I have been at the Census for, it'll be 18 years in June, and I started my career as, in 2007, so luckily the year before the job market was scarce, so it was a really, really good time to get into Census, but I came in at an entry-level position.

I had a degree in sociology and a background, minor in statistics, and the Census Bureau does require 15 credits of statistics to join. So, if you're interested in coming to Census and you're in college right now, definitely plus up on some of those social stats classes for the future.

So, I came in just doing, answering phone calls from the public, and I was very fortunate that I also got to start a criminal justice master's degree at that same time. It was before virtual, so I'd spent all my day in D.C. working, go to Baltimore all night, and go to school, but it worked out really well because as soon as there was an opening in the criminal justice branch at Census, I was able to join, and that was really where my ability to dive into data, surveys, the survey responses started to occur, and that was about maybe 2010.

So, there still wasn't an official data science position at Census. That is new as of, say, the 2020-2021 is when they started hiring data scientists. So, all of us survey statisticians were doing data science in our own way, and that might involve SAS. There was a few SPSS licenses, but primarily SAS was the way we were analyzing data, diving in, creating visuals, some pretty cool deliverables, but still not the level of the things that we see today.

So, fast forward to 2017, and I switched to another criminal justice branch. So, we have two at Census, and that was where I got to really dive into the data science field. We were offering a self-led data science training program in Coursera, and that was in both R and Python. So, I started with the R program, and I have not looked back since, and it has really changed the ability to look at our data, to touch the data, to correct data, to link data across time, and I have been a huge proponent of that at Census because we're all doing some amount of data work, data science, in our respective surveys.

So, that led me to, yeah, the R field. So, a little bit of what I like to do for fun outside of work is I play the piano. I spend a lot of time outside exploring in the forest that I live next to, and I also have two kids, an elementary schooler and a teenager. So, all the fun that they're getting into now, I get to join along for the ride. That's what I do outside of work.

Awesome, and I feel like that is a good, you reminded me of maybe a potential fun fact for me, since you play piano, is I did classical voice all through school. So, I had to learn piano, and I hated it. I'm really bad at it. I was much more interested in the classical voice part than having to learn piano part. Well, I was hoping that you could give us, because we are all from totally different industries, a little bit of a slice of an example, right? So, what kind of problems does your team solve? What type of data do you use? And we know that you use R and or Python. Maybe some insight into packages that are really important to you would help too.

Criminal justice data work at the Census Bureau

Absolutely. So, currently, and I will talk about my experience in the Criminal Justice Branch, but you can extrapolate this experience to all different programs all across the agency. We collect data on about 300 different topics. We have surveys. We collect administrative records. So, my experience is very similar, I feel, to other people's. It's just everyone has a different part of the survey life cycle that they're working on.

So, within my Criminal Justice Branch, we collect data from facilities. So, we collect juveniles that are in residential facilities, and then particular information on inmates. So, this is going to a central reporter that's giving data on behalf of their population. So, it's not always, I would say, sometimes it needs a little bit of correction when it comes back, and R gives me that ability to highlight what's an outlier, what's missing, what needs to be flagged. Automated workflows that we can run our data through to highlight where's the problems and how can we fix them.

Another really neat thing that we have started doing in the last few years is linking our data, not only across time, but across programs. We have four separate surveys, and before, we were really looking at snapshots in time. What did the survey data look like for a year? And then we would pass it off to our sponsor for their own analysis, and they would create the publications. So now, I've been able to now link data for 20 years for one survey, 14 years for another survey, and then my survey has actually been collected at Census since 1930. So, I've been able to do a little bit. I know it's one of the smallest surveys at Census, but the longest standing surveys at Census. So, it's pretty cool to look back in time with it, and these are abilities that I did not have before and that we did not have available in our branch.

Now, because I started in this field, all of us are starting to, in the Criminal Justice branch, we're starting our own data science learning journey. We have people learning Python, we have people learning R because both of them give us the ability to work with the data on such a deeper level than we ever have been able to. So, it's like we're trying to take the best of both and apply it to our work in the most incredible way possible. Like I said, we have some deliverables that we've never been able to produce before, all because we can now link across time, link across programs, and even that administrative data records that we're able to bring in is now giving us supplemental information that we never had access to.

Transitioning from SAS to open source tools

That's awesome. It makes me want to ask you a million questions about the logistics of an organizational transition from SAS to R and Python, because I don't think you're alone in that. I think that there are probably a lot of people here who could raise their hand and say, like, yep, I'm in a heavily SAS org, or I was at one point, and we're trying to move into the open source world. So, when I first talked to you, Jessica, I was just really, really impressed by the internal support system that I heard you talk about at the Census Bureau, because you were mentioning things like training programs and workshops and resources for using the Census API, and it really just seemed to reflect this broader agency commitment to modernizing data practices that is so rare and great. So, I'd like you to talk just a little bit about that. I'm seeing in the chat many people agreeing with the SAS experience here.

So, any advice, tips? How did this happen? You know, I think that Census was one of those really incredible examples of where the people doing the work and then the high up came together to see that this was going to be a valuable transition for everyone. So, like I said, I started at Census long before we had any of these. I mean, we had R, and we had people using R, but it was very limited to the geographers, the people who were mapping, the people who were using our sheet files. Oh, really? Yeah, it wasn't very, and I only know this because I've asked people who only know R and not RStudio, and they say, oh, I've been using R at Census for a long time, and they like their base R. They like creating the functions themselves, not using the packages. So, that has given me a sense of how it's evolved since the beginning of R at Census to now.

But anyway, so I came, and I, like I said, I asked for an SPSS license. They told me that it was really SAS. It was SAS or nothing. So, I took a bunch of SAS classes. I never got the hang of it, not the way that the people who are really SAS savvy were able to.

So, and that was not only my experience. It was many other people's experiences, and they, that movement of, you know, there has to be something better. There has to be a better way to work with the data brought the staff together at that analyst level. We helped each other. We didn't need our bosses to tell us to learn this. We just knew it was going to be valuable. We were working with too big of data, too disparate of data to just rely on something that we only had this tentative grasp on how to use.

So, when we started creating, after the training program, when we started creating deliverables, showing them to our managers, and that's when the buy-in really started. That's when people started seeing, you know, these are things that we can show to our sponsors. These are things that we can publish on our website to show the public that we're aligning with the way the world, I guess, the education system, really. In college, people are teaching R and Python. They're not teaching SAS as much. I even heard the college that I went to no longer uses SPSS because it's too expensive. I think they're using a free spreadsheeting software. So, even that SPSS is gone.

So, when we started creating, after the training program, when we started creating deliverables, showing them to our managers, and that's when the buy-in really started. That's when people started seeing, you know, these are things that we can show to our sponsors. These are things that we can publish on our website to show the public that we're aligning with the way the world, I guess, the education system, really.

Oh, wow. I know. I was surprised. So, that, yeah, that show-and-tell approach, I think, is where that buy-in started. And then, once managers started saying, you know, not only do we like this, but we want to add to the skill set. We want more data scientists and we want more people taking data science training. That's when we started seeing that top-down. And then, we met in the middle. And now, it does seem at every level, there's a value of data science. You might not understand data science as a manager, but you understand how you can leverage it. And I think that's gone a long way towards encouraging new deliverables, new products, innovation, the things that before we were scared to do. Now, we're encouraged to try it and then show others and then see if we can bring them along for the ride. So, it's been a complete change from when I started to now.

That's awesome. Well, hopefully, in a little bit, you can talk a bit about maybe some of the conversations that went on early because I know you've been there for a long time. And that brings me to Kylie's question. Kylie, do you want to hop in and ask? Yeah. Hi, Jessica. My question was about retention. So, you've been at Census Bureau since 2007, so almost 20 years. And that's

quite rare these days. People hop around. And I can understand that sometimes in government, there are bureaucracies that are annoying. I work for a foreign government. But I just wanted to know why have you stayed and what is it about your job or your organization that has contributed to you staying at the Census Bureau? That's a fantastic question. And, you know, my plan wasn't at the very beginning, this wasn't my plan. I really wanted to work in a criminal justice space.

When I was little, I had dreams of being in the FBI one day. But, you know, life has a different path. And Census, I believe, as time went on, as they started saying, let's try to innovate, let's not stick with these traditional tables, these traditional ways of delivering data that we were all used to that are a little bit harder, I think, for an average analyst to be able to just work with. But now that we have the right tools, I think that those of us who are really, really, not only enjoying the data science space, but wanting to see how far we can push it, it's made it an attractive agency to stay at. I think that if we did get a lot of pushback, like even if now my bosses were saying, you know, I know this is really cool, but we really need to go back to the old way. I'm not sure if I would be as excited to stick with it. But the Census has truly changed their mindset and I feel like pioneered this modernization effort. And I think rather than considering a job, I almost think like, where are we going to be next year? Like, I'm excited to stay and see what does the next year look like? What does it look like when we're all in the cloud? What does it look like when I get to move to the cloud? So these are the things I'm excited for rather than just that day-to-day, oh, I have to get up and go to my job. It makes it fun.

But the Census has truly changed their mindset and I feel like pioneered this modernization effort. And I think rather than considering a job, I almost think like, where are we going to be next year? Like, I'm excited to stay and see what does the next year look like? What does it look like when we're all in the cloud? What does it look like when I get to move to the cloud? So these are the things I'm excited for rather than just that day-to-day, oh, I have to get up and go to my job. It makes it fun.

Awesome. Thank you for the amazing question, Kylie. And I see lots of really good questions rolling in. Don't forget that you can vote in Slido. So if you go peruse those questions, vote your favorite one. The top of the pile right now is Abigail. I have a asterisk next to yours. It looks like you might not be able to ask right now. So I will ask Jessica, what does your hiring process look like? What technical skills do you look for versus train people in in your department?

Hiring and resume advice

That's a great question. So I have only gotten to be part of hiring one time, and that was before I became an official data scientist because I had a data science skill set. They allowed me to kind of come in and help assess skill set. When something's new, you don't really know what you're looking for. That's kind of where our HR was at. So that was the one time that gave me the ability to see, OK, how are managers evaluating these resumes? I believe strongly, and this is going to go into my what is my advice at the end, but I do believe strongly that the public get hubs with practice examples, data science pipelines, projects that look similar to the agency or the organization that you're applying to go a long way, more so than listing out the packages, the skills. Those are very important to put on your resume. But I remember from my experience just diving in and being able to see a project from beginning to end, help me understand what value of contribution that employee would come in and be able to apply towards our projects.

So every government hiring goes through USAJOBS, and I do believe that that might be a little scarce right now, but please don't give up. I believe that hiring will come back soon. And that's where that first step of resume. And actually, I think that the military might still be hiring for data scientists. So still a good advice for now. And that's the first hiring managers, I believe, are starting to bring in people who have that data science experience to kind of go through the resume. But if not, I do think it's helpful to tailor it towards someone who might not fully understand, like they might see, I'll use TidyCensus, for example. That's one of our favorite packages here. So they might see the TidyCensus package, but they might not understand, well, what does that do for me as a data scientist? I know that gives you all the census data at the tip of your fingers. If you can use TidyCensus or the Census API, you probably have a very good skill set. Without being able to see that, I guess, through the GitHub, through explaining projects through the pipelines, it's hard to understand really how meaningful and truly impressive that skill is. So I do believe that as people are applying for data science jobs, especially in the government, the more people who are

That's amazing. I love the TidyCensus package, and we have a TidyCensus fan club forming in the chat right now. A lot of us use it. That's a package, in case anybody does not know, by Kyle Walker, right? And if anybody has the link and wants to put it in the chat, for anyone who doesn't know it, please go see.

Most surprising findings in the data

All right, let's see what we've got next on our question list. Don't forget, y'all can go in there in Slido and upvote. I also see some people have added their names to their questions. Thank you so much. That's amazing. I have an anonymous question that is the most popular one next, and it says, what's the most surprising result you have come across in your experience?

I'm thinking of a very recent example. I have uncovered some very interesting things, but not even myself. I've had other colleagues come to me and say, this is this really interesting gap or hole or piece of information I've found, but for me personally, one of the most interesting things that we recently saw was how data can kind of conflict with the news. The data that we're being provided is not the same as the data that we're seeing reported in news articles, and we're able to now more or less sync up information with the information that we're seeing in the public, and I find that to be really interesting sometimes, where it either matches, where we can say, oh, yeah, this is what was reported to us, this is what I've seen reflected, or, you know, this is what was reported to us, but that's not what I've seen reflected in this piece of information I've found, and I feel like that level of understanding these, again, large data sets, you really need to analyze them and aggregate them to see where some of those holes are, and that, I feel, is probably one of the more fun pieces of data science. It's the what did you find that you either didn't expect or is missing or you know you can correct or should correct from your analysis. So that was the, again, the recent one that I was like, wait a minute, I just read this article, but that's not what the data is telling me.

Interesting. I feel like in school, a lot of us learn about the Simpson paradox, right, where you have data that is showing you one thing inside of categories, but then when you aggregate it, you see a conflicting pattern in their data relationships, right? I think that that's really, really possible with people data, census data, but you only get to see that if you get the whole picture, so you're in a really enviable position of having so much data.

Census data shaping policy and real-world impact

All right, let's take a look at our next fabulous question. There is no asterisk, so let me see who this is. Jeff, Jeff, would you like to unmute and ask your question live? Sure, there he is. So I was curious, I've played around a little bit with the Census Bureau data, and I think it's an extraordinary source of insight. I was wondering if you had some examples of how that data has been used to shape public policy.

You know, that is something that I can't answer. I think that I can answer it with a non-answer. So as Census, we really try to stay away from that policy. I work in a reimbursable area, I work in a reimbursable area, so my responsibility is to collect the data and then pass it off to my sponsor, the Bureau of Justice Statistics, to create their own analysis that will then shape policy. There are areas at Census that are more directly involved with that, and like I would say ACS, the decennial, of course, but because I'm so far removed from those areas, it wouldn't be something that I could speak on. But I do think it's interesting when I find census information just in the general news, and I think, oh, you know, I saw an email about a seminar months ago where they presented that as a first peak, and you can kind of see where the beginnings of projects end up to become real solutions. A good, maybe a good example where it didn't shape policy, but it was that real-life impact is during the hurricane season, we were, had some of our colleagues were working with FEMA, and they were able to use the census data to help address the community resilience estimates and which communities were going to need the most help, which communities were in danger of failing or, what's that word, not being able to rebound because they didn't have all the resources they needed. And I always like this, where census can actually, real-time help, we're so used to collecting data and then maybe it's another year or so before it actually, something gets to happen with it, but those community resilience estimates, those emergency situations is when you really get to see quick thinking or census data help people in the long run, and those are my, some of my favorite experiences.

Awesome, yeah, I think that that's really, really important to realize is that all of that data, even if it's not going towards something that's being used immediately, how lucky we are to have it there, to just have it there so that when something does happen, we have the access to it and we can know what's going on and we can use it to respond to things. Thank you for the, oh, we have somebody unmuted, sorry, we'll fix that. Jeff, thank you so much for that great question.

Open-source tools for geospatial work

Next on my list here, what do I have for questions? Noor, I have a question for you, are you available to ask that? Hello, hello. Can people hear me? Yep, we can hear you, how are you? As long as I don't think about how I'm unemployed, I'm great. Anyway,

so I have a question actually that kind of intersects open source and not open source. Along with SAS, there's things like ArcGIS, and I'm curious about one, how does, since, how do you, like, since we're dealing with the geospatial aspect of the data, as well as what are some good R and Python packages for people who are trying to, like, break free of the ArcGIS criminally expensive infrastructure, basically?

That's a fantastic question. You know, since this was an open source conversation, I had intentionally left it to open source, but I am Tableau certified and have taken a full year of ArcGIS training, so those are two tools that we use and we rely on heavily. Similar to what you just said, how do we break free from these contracts, these proprietary software, we do want to try to loosen our footprint and our reliance on these softwares, but that being said, our geography division does use ArcGIS probably more than any other area. We also have the, what's called a community explorer, ArcGIS community explorer, it's kind of a web-enabled place where we can build dashboards, we can give little surveys, and we can give more interactive dashboards to our teams. So, again, these are used very, very, across the agency, they're used by some of our higher level, the people who are creating the publications, the things that are going out to the public. If you are not producing deliverables, maps, dashboards that are going to the public, you are encouraged to use R or Python for your solutions. Keep them internal, don't use a license if you don't need it, but if you are publishing, we currently don't have a way to publish outside of ArcGIS and Tableau. So, we're kind of in that in-between space where we haven't been able to move fully past them, we do still have, like I said, I believe our ArcGIS footprint is pretty heavy, but we know that these are things that we would like to do less of.

That being said, we have a presentation next month on using the R and the ArcGIS bridge within our internal agency, so that tells me it's not going away anytime soon. And nor would we want it to. I think, again, it's a very valuable tool. It's, in my opinion, I had a hard time orienting myself to ArcGIS. I thought it was very complex, but Tableau, I always will have a love of Tableau. It's so much easier to get your, like, just quick findings, right? But we are trying to, within our own user groups, push people towards visualization and mapping with R and Python. So, you ask the packages, Tigris, which is a compliment to TidyCensus. Tigris will give you all of the census shape files. I find that to be a fantastic package. I don't have a Python equivalent except for this censusdisc, which is a new package in Python. Someone will know it and put it in the chat, I'm sure, and that might have a mapping component, too. But the Tigris, I think, is amazing, and I know I'm supposed to stay away from the chat, but I really would love to put this three-hour webinar series in.

So, that's like a nine-hour webinar that Kyle Walker does every year. I find it to be the most, one of the most valuable resources to using TidyCensus and Tigris. The content doesn't change that much, but the practice and when they add new data sets, he gives a nice overview of how to access that new data. So, that is, I would truly recommend looking at that if you're interested in mapping with census data, and again, using all of the publicly available data sets, shape files that we already published to our census website. It's all at everyone's fingertips. It's just understanding how to use it. That's sometimes the hardest part.

And I'm actually going to put a few more links in the chat if that's okay. Sure. Yeah. We tell our guests to avoid the chat just to not get confused because we all know that our chat is an absolute party and that it can be hard to keep up with. Somehow Lily has a superpower where she can chat and hear an answer and respond. I don't know how. I think it's because I was raised by AOL. I lived on chat rooms and stuff when I was a teenager.

All right. I have another great question from one of our amazing regulars, Arsenis. Would you like to unmute and ask yours? Sure. Hi, Jessica. Thank you for joining us. My question is about data sharing across agencies. And I'm not asking at the level of policy. I'm

asking how you deal with challenges in data sharing, such as administrative data at the command line or when you're writing your code. What are some of the statistical methods? That's what I'm asking about that you try to employ when you have a gap in data that you know exists, but you can't get to. And are there any methods that you've used and have had success with?

That's a great question. So I personally have only done a few of those little linkage projects, but the big ones that we're working on, and we actually just had a presentation on this, is the fast link package and probabilistic linking. And that we've had a few demonstrations on and it always seems to be not only very popular, but in the chat, kind of like you said, when people are presenting in the chat, but I see people light up and they're like, oh yeah, if I'm doing linkage, that's what I'm working on too. But it is a lot harder because we have this linkage process that randomizes people. It groups people in like buckets, whether that's facilities or peoples, and that's how we can do this linkage across different spaces. So sometimes it's not a one for one, especially because we have to protect the data within our program.

So you had mentioned sharing data outside, much harder to share data outside the agency. Even with myself as someone who works on reimbursable surveys, there's policies on when I can share the data, who I can share the data with. But internally, if we're working, as long as we have agreements, and these are contracts, these are real agreements, this isn't just like a verbal, hey, let me use your data. We're very, very big on data protection, data disclosure. We have training titles on it. But if we are doing it, we do it more with a random approach where we're not giving away the facilities that I'm working with or the inmates that I'm working with more, what are the people that fit those similar profiles and how can we impute data from there?

And that's another thing that we're doing is we have one census product within our branch, and it's a census of people. And so if we don't get that full census, then we will do imputation hot deck. What are some of the other ones? The hot deck, the cold deck? No, maybe not the cold deck. I can't, but there are a few different, like three different methods on how they will actually impute to get the best information possible. Awesome. Thank you for that fantastic question, Arsenis. Allen, you had a question that I feel like is really similar. Do you feel like yours got answered or would you like to add on to that?

I think it's really close. Yeah, Jessica. To Arsenis, you referred to some of those regulations around sharing and confidentiality and stuff. And so maybe I can generalize a little to like how much does that play in your day to day? You obviously need to be aware of those things. Like when data comes to you, is de-identification and stuff the first step or is that handled somewhere upstream? Sort of how day to day do you have to be conscious of those kinds of things to make sure that you're meeting all of those obligations? Maybe just a little bit more general from your response to Arsenis. Thanks.

Data security and open source at Census

That's a fantastic question. And this is also very relevant to how we use open source at Census. So our Census network is incredibly locked down. Not as locked down as like probably DOD or some of the other agencies, but even I've heard other statistical agencies, maybe an agency like NOAA, where they don't have those same levels of restrictions. But we have a very, very strong firewall. We are not allowed to hook our laptops up off of network. And a few other little precautions. But all of that leads to the idea that if you're working within your computer, your server, so I have my own criminal justice server that our team works on that we don't have to worry as much about that. Like we're not sharing information outside of our network. We're not sharing it with people outside of our team. We have this policy called need to know. And every year we get trained on it, that if someone needs to know, because it's related to their work, you may work with them on it. But if not, it is illegal for you to disclose that. So I wouldn't go to even that second criminal justice branch and be able to talk about this, even though we share the same sponsor, it would be a violation of my agreement.

So that does mean that there are things with R and Python open source that we can't do, because we don't want to risk sharing, inadvertently sharing information, like you said. So I don't, I don't and not many people, areas that I know of have to do any anonymization, any sanitization as it comes in. It's more once it leaves your secure network, it must be in that proper format to be able to share. We can't use AI with this information. We can't use packages that are off of CRAN, or PyPy, I believe is the equivalent. So that's some of our security procedures to make sure that this information doesn't get in the wrong hands.

We can't use packages that are off of CRAN, or PyPy, I believe is the equivalent. So that's some of our security procedures to make sure that this information doesn't get in the wrong hands.

And again, that secure server, we have a secure server because we work with individual people, but not everyone has that, like if you don't work with that secure data. So those are extra steps that you, we take, or an area will take to just make sure that the information stays within the area. Well, thanks. I appreciate hearing how careful and thoughtful all of that is with respect to that. Thank you. Very. We take that the public trust and the trust that you all, well then it's public trust, and the trust that we're not going to disclose anything. That is probably the number one oath that we take as census public servants.

Survey weights and open source conversion

Thank you. Awesome question, Alan. I have an anonymous question at the top of my, my list here. Don't forget, you can go hop into Slido and vote on these questions. The voting has been really fun and really helpful for me. So it says, attended a talk recently about survey weights, not just people weights, but geography weights. Any resources you look to, you look for, for applying weights correctly? This is a very specific and technical question. If you can't answer it, answer it in whatever way you feel you can. In the most incredible coincidence or fate, as fate would have it, this actually, that was my last R project that I finished yesterday, is I had to create sample weights for the sampling file for our upcoming collection.

So, and it was a challenge. We had to deal with some amount of double coding the way it used to be in SAS. So that's something that I work on. I try to convert some of our SAS programs to R. So I did sampling, weighting, and now coefficients of variation, which are still eluding me. So if you had asked about that, I'd say, I don't know. But the sample weights, that was really fun. And we don't do it by region, we do it by strata. So that is actually region for one of them, but that could be the size, that could be if there are a secure facility versus a non-secure facility, size, region, and then maybe a few other details. Are they male only, female only? So that was the challenge, right? Like making sure that the survey weights fit the expected sample versus the population. And it all matched yesterday and it was amazing.

But I don't know if I would have been as successful if I did not have the SAS output to compare to. So it was really not me being able to validate it because I'm not a math stat. I am familiar with statistics, but I'm a social data scientist is how I look at myself. But I can use open source really well. I can understand statistics. So I understood that my output made sense, but if I wanted my sponsor to buy into it, it was still going to have to match the output from years prior. So there is some level of that, that it's not just validating it, especially when we're moving from tried and true processes in SAS into new processes in open source. There's that extra level of, okay, it's either got to match or you have to give me a good reason why it doesn't match, which we've had that too. And then they can accept the results going forward.

Okay. I'm not sure if that answered it, but that was, like I said, just yesterday I finished creating the weights. I think that's a great answer. And I really appreciate it. I have a couple of questions that were more technical, so maybe those will pop up as well. But we have a great one at the top of the voting list right now. It's an anonymous question, and it's what professional development do you think newer data analysts should seek out?

Professional development for data analysts

So this one's a little bit more broadened. I love training. So training is one of the things that I find not only to be the most important as someone who's trying to maintain skills, learn new skills, but as myself, leader of our user group at our agency that I try to promote upskilling opportunities. And not only the ability to go back to skills that we all maybe know, or we don't use as often, so we forget, but to be able to adopt new skills too. And I do think that that continuing education, the continuing of showing that you are still learning, you're taking training, you are furthering your professional development, does go a long way.

That being said, I've learned that in the government level, I learned this through my own experience actually when I was hiring, that some of that professional development doesn't count towards an official education requirement. So I do, like I said, I think that training is the most valuable thing that someone can do for themselves, whether that's we use LinkedIn Learning, Udemy, Percipio, Pluralsight, Coursera, all sorts of great options, Datacamp, Posit Academy, I believe you guys have, all sorts of great options. But if there are certificates, like I know I'm hoping to be able to take the master's certificate in data science at our local college. I think those things go a long way towards hiring. So it's not only is it the professional development that you can do in your day to day, but if you don't have a data science, if you're maybe at that, like I was at that survey statistician level, I had to show that I had the qualifications to move into a data science field. And that was luckily through grad school, I got enough math credits.

Awesome. Yeah, I think that's an interesting distinction between survey statistician and data scientist. I think a lot of people who are in survey stats positions probably do a lot of data science already. Do you feel like you're maybe already doing a lot of the stuff that you ended up doing? I think that all, I don't know if this has been a popular opinion or not, but I've been telling people this, I believe that all surveys... Uh oh, we just lost you. We just lost your sound. Oh wait, did anybody else lose their sound? Yeah, it might be her headphones if they died.

Maybe, wherever your mic is attached to, Jessica, but you can switch your mic in Zoom. The little button next to audio, this little up button, and change your, select a different microphone.

Okay, now I might have to talk. No, it can't hear you. Let's see, in the bottom left hand corner, if you select up on audio, you might need to just switch to your computer. Speakers instead of the headphone.

No, you sound very distant and far away, so maybe... Is it your microphone that's behind you on your headset? Yep. Can you hear me now? Yep. Okay, so that's... Sorry about that. That's all right, go ahead.

Survey statisticians as data scientists

Okay, and the question was survey statistician versus data scientist. Yeah, yeah. I do believe that every survey statistician is essentially the beginning of a data scientist. You just might not have the right tools, or you might not have gotten access to the right applications to be able to take it to that next level of true data science that you can show to your managers, but we're all working with complex data. I was trying to do the best I could with data science in Excel, so even... And I do have someone on my team, actually, who probably uses Excel more like magic sometimes. I look at it, I'm like, this is true magic. This is data science at its own level.

So I do believe, like I said, the survey statisticians at that fundamental level, you're working with the survey data, you have the ability to take that next step into the data science role. Where we're limited, I would say in the government, is that you have to go from 15 credits of math to 30 credits of math, and not everyone has that, because you probably wouldn't have gotten that unless you had a specific math major. So I think that's more the limit of bringing people into this new field, not the lack of ability or the lack of even working with the data, just that math credit.

I do believe that every survey statistician is essentially the beginning of a data scientist. You just might not have the right tools, or you might not have gotten access to the right applications to be able to take it to that next level of true data science that you can show to your managers, but we're all working with complex data.

Yes, I believe that everything I was doing in my old survey stat role was some amount of data science, it's just data science, just not with the right tool.

Okay, that makes perfect sense. I have a Slido question at the top of my list here from Shivani. Shivani, would you like to unmute and ask your question?

Yeah, just give me a sec. Which of your work are you most proud of?

Oh, you're asking me what work I'm the most proud of? Yeah.

Last year, or maybe it was two years at this point, I got to create my first publicly available Tableau dashboard. It's the only thing I've ever been able to publish. It had nothing to do with my survey. It was something that I got to come in and as a resource share, because they needed a Tableau developer. And like I said, I have that certification. So it's the only public piece of data science that I've been able to get put up on the web, but I'm very proud of it. So that's a great question. I love that question. I love those types of questions. Thank you so much for asking. We have another Kylie question. Kylie, you want to hop in and ask that one?

Quality control and reproducibility

Yes, sure. I have to remember what my question was. Oh, yeah. So kind of going back to what you're discussing a few questions ago, when you were talking about doing the survey weights, and I just wondered, do you have any quality control practices to make sure that the weights that you actually calculate are correct, and also reproducible research practices? So like, do you do things so that your colleague down the hall could run your code and get the same answer?

Those are great questions. And I believe I thought later that there was one part of your question I didn't answer earlier about the turnover, and we did have a significant amount of turnover in 2020. So you can see people that were hired since 2020, and then the people like me who stuck around from before 2020, but that was a huge divide. But okay, to answer your question, we do have a quality control process, very, very strict review. So I am in a survey statistician type role as a data scientist, but that does not make me a

mathematical statistician. So let's go back to the weights that I produced. That will have to be reviewed by the math stats, and then I think it goes all the way up through to the division chief, and then they sign off on it and say, yes, this is mathematically sound. We agree with this.

But the math stats are mostly SAS. I feel like math stat and SAS really, really went hand in hand for years and years, and now we're trying to break free of that mold, but that's where that double coding is coming in handy, is the math stats are doing their process, running it the same way. We really didn't want to do double coding, to be completely honest, but we've learned you can't not do double coding if you want your sponsor, and during that review process, something to compare it to. So the people like myself, the data scientists, the people, the analysts are trying to convert the processes so that the mathematical statisticians can pick them up for years to come.

Version control and GitHub adoption

Now, how do they do that? Version control, GitHub. So we are huge on GitHub, GitLab. We're actively making strides to get the whole agency enabled with GitHub. Right now, it depends on what area you're in. If you have, again, GitLab for some areas, GitHub for some others. I use GitLab, but I can't share things with people who don't have it, so we're trying to get to that point that we can not only share within our agency, but my sponsor, I had to email her all the code files. She's very R-savvy, so she knew as soon as I gave her the code, she understood the process.

What I'd really like to give her is the GitHub or the GitLab, but right now, everything's internal, but that is our really big push this year is to get everyone enabled, start practicing version control, so that, yeah, we don't lose these really, really important code files. And a really cool thing that we're doing is sharing, like you said, sharing those best practices. So last year, I converted the sampling process from SAS to R, and, oh, so there were two processes that were just a little bit more complex than I was used to at that point. I was able to borrow the code for one of our larger surveys. I was already out in the field, had already been vetted and approved by many people, and it was the same sampling process. So I used the code that they had put on their GitLab, and then their sponsor knew, okay, this is not something that Jessica just made up. This is something that's already been put into practice, already shows that it works well, and we're trying to do more of that. We're trying to share rather than recreate.

We're all doing, I would say, a similar piece of different, a similar process for different surveys. So if we can get away from creating silos, get away from that redundancy, and all align and modernize together with similar code, I think it's just going to make for a much better way to not only get these things out to the public, but make sure that for years to come, we have consistency, and we are putting together things that link over time, and giving the public information to really understand the data that they're looking at.

Awesome. Wow, thanks. Can I ask a follow-up, Libby? Is that allowed? Sure, real quick. Okay. Well, I was just wondering, you say we want to try to prevent silos and share resources because people are doing similar things, but how do you do that? Because we're also trying to do that where I work, but that's hard. Is that working groups? Is that documentation? How do you know that someone is doing a similar thing if you don't know?

Technical support groups and upskilling

That is a great question, and that's really where our tech... We don't have user groups, we have technical support groups. So we have an R technical support group, and I've been leading that since 2021, and then we have a Python technical support group. And we have membership from all over the agency. It's open to anyone who's interested, and that's staff, that's contractors, that's people in our regional offices, and that's kind of cool when we have people come in that are not necessarily at headquarters, but they're like, you know, I really want to learn these mission-critical skills. And that's where we can give that exposure, that show and tell, how do we share best practices? An example was not last week, but the week before we had a presentation on tidy data. Fundamental, but so important for so many people. And a lot of people came and they said, you know, we really need to think about getting our data into a tidy format before we move into the cloud, before we start putting...

One of the biggest challenges of the census API is that you, not tidy census, but just the census API, is that the data's in different formats. You have to do a lot of wrangling to get it into a similar format so you can combine it. But maybe we could do that before it even gets posted to the API. When we move to the cloud, if we can put all of our data in a tidy data format, then we'll all be in a better place to share, combine, link over time. So it's those little things through these technical support groups where we spread information, we give upskilling opportunities. Next Thursday, we're doing our R-SNAC session, where we're going to do just the basics of reading in, writing out data, and creating synthetic data. And how do the different file formats take up, like, what are you losing with different file formats and what are you gaining in space? So just those little things as we all align. And if we can help people think of better ways to do it, better ways to store their data, process data, analyze their data, then again, I think that those silos will slowly start to break down and we'll see that we've made big strides together. Awesome. Thank you.

Career advice

Thank you. Thanks for the great questions today, Kylie, keeping us rolling. And I wanted to hop in with a question that we try to ask everybody because it's always so great, which is, do you have any career advice that was impactful for you or that you'd like to share? You know, I, because I did not start as a data scientist, I didn't get what I think would be the best advice for now. And again, I think that's those practical projects, the projects that you can put together, the domain, the ability to get that domain knowledge, I think is so much more important sometimes than the data science skills themselves. I've met a lot of people with some really, really great skills and that can innovate and modernize, but if they don't understand what they're trying to accomplish, it's very difficult, lots of roadblocks you can only get so far.

So that just demonstrating a subject matter expertise in the field that you're trying to get in and showing those projects and being able to talk to a recruiter, an interviewer about the data or the projects or the data science that you're most interested in and how it applies to the space that you're applying to, I think really does go a long way. I have never not talked to someone who has a love of data science and not been like, you know, I really would love to add that person to my team. They know what they're talking about and they're interested in this field.

So that just demonstrating a subject matter expertise in the field that you're trying to get in and showing those projects and being able to talk to a recruiter, an interviewer about the data or the projects or the data science that you're most interested in and how it applies to the space that you're applying to, I think really does go a long way.

Working with large datasets

Thank you. And you know, I've got a couple of questions left here. We have about five minutes left, but I bet we could squeeze in a couple more, maybe one more. So the next one is anonymous and it says, how do you deal with large data sets? That's a really broad question, but I think it's interesting. Like how big is your data? No, that's a great question. My data is not big, but I feel this question all the time because I have to help other people. People are wondering, how do I work my large data set?

So we have a space called the IRE. It's our Integrated Research Environment. And that's where a really, really, really large data is housed. And that's where we have a few people that have tips and tricks on parallel processing and the data table package and different things that will speed up your analysis. But that being said, that is probably one of our biggest challenges. And I believe that our move to the cloud will kind of speed up what we can do with the large data. But yeah, I've had some people come to me with some big data challenges, especially every summer, we'd love to do Civic Digital Fellows. I don't know when we'll get to resume that program, but if you are interested in working in the government, this coding it forward, Civic Digital Fellow, I cannot recommend it enough. Some of the most impressive people I've worked with have come to us from that program and stick with us. We do really like to bring people on who complete that program.

They always struggle with the big data aspect. I had one person, I think it was halfway through, and they're like, I just need to subset it into like 20 different files and work with it that way. So it's an evolving challenge, but one we're really hoping, like I said, in the cloud is going to be less of a problem.

Resources for understanding census data

Awesome. I have another one that maybe you can answer really quickly with the few minutes we have left. It's anonymous. And it's, do you have any tips or resources for understanding census data? Because it's out there, we can use it. There's a lot of fields, the question says, and even data sets like the ACS, Decennial, and PUMS, P-U-M-S.

I am putting this link in the chat, and that's our guide to working with census data. So there are, it's, you know, we're really working to build it out with real life examples, but right now, it at least gives that preliminary guide on how to work with census data. That being said, it is confusing. And you know what, as, it's almost embarrassing to say, but even though I worked at census for so long, it wasn't until 2021 when I attended Kyle Walker's webinar, and that's why I recommend it to everyone. That's what helped me understand census data. I was so siloed in my own criminal justice data that I never was able to understand until I attended Kyle Walker's seminar. So I would definitely look into that. He has a tidy census book that's free, or analyzing census data with tidy census book that's online and free, and just practicing it.

And then if you master tidy census in Tigris for visualization, I would suggest moving to the census API package and trying, not necessarily trying to get different data, but seeing how you have to query and call the different data sets using the census API. And that kind of gives us like the start of the understanding on how to access the census data through, again, the census API, it's all the data is a little bit in a different format. So you have to really start to understand nuances, reading documentation, but once you get the general hang of it, it does seem to fall into place. Awesome. Perfect. These questions on Slido are so good. Y'all who are putting anonymous questions out there, I encourage you to put your name on them so that I can reach out to you and thank you for your amazing question, because they're great. So we have a few minutes left, and I want to see if I can sneak one more in.

Foundational math for data science

What are some foundational math courses that you'd recommend for someone looking to upskill? For example, calculus, linear algebra, probability theory, etc. The linear algebra was probably one of the best that I had. I agree. Yeah, because just I feel like it helps you even just visualize what you're trying to accomplish. But that's honestly one of the places that I frequently am trying to upskill and focus my training is math for data science. So on LinkedIn Learning and Udemy, I found a few courses that are like statistics specifically for data science, math for data science, and that seems to kind of help subset. I mean, math is just such a huge field. If you dive into it, you don't know if you're applying it correctly, but I do believe the ones there are good courses geared towards that specific data science math.

There are, and there are also some really great YouTube channels. Noor just shared in the chat, StatQuest is super helpful. Yep, Josh Starmer from StatQuest is great. There are a couple more fabulous YouTube channels that are blanking. Three brown, one blue, maybe is that one that I'm thinking of that are really helpful. I'm just going to echo like linear algebra. Yes, it's going to help you understand things better, especially if you work with things like PCA, principal component analysis, like understanding the way that vectors work. If you use vectors all the time anyway and you don't understand vectors or eigenvectors, yeah, go do linear algebra.

All right, so we have a minute left. It is time for me to say thank you to Jessica. Thank you so much for everybody for joining us. The chat has been so full of resources. I'm going to try to get some of these together and post them on LinkedIn or something. I want to remind everybody to join us next week. Jeroen Janssens is going to be sharing his wonderful experiences with us. He's a senior developer relations engineer here at Posit. He's also the author of Python Polars, the Definitive Guide, and Data Science at the Command Line. Fantastic person and wonderful stories to share. Let me put his LinkedIn in the chat so you can go check him out.

In the meantime, all right, everybody, thank you so much. See you next week on Thursday. Have a wonderful day.