Transcript

(Editor's note: transcripts don't do talks justice. This transcript is useful for searching and reference, but we recommend watching the video rather than reading the transcript alone! For a reader of typical speed, reading this will take 15% less time than watching the video, but you'll miss out on body language and the speaker's slides!)

[APPLAUSE] Well, hello everyone. Thank you so much. It's so incredible to be here today at the Deconstruct Conference. I'm having the time of my life. I'm super excited. And more importantly, I am here to welcome each and every one of you to The Odyssey, Lessons Learned from Learning.

Now I'm sure you all probably are familiar with the concept of an Odyssey. OK, I'm going to take you on a journey, adventure, great. You know, you might not have signed up for it, but it's going to be real fun. We're going to all have a blast. But Lessons Learned from Learning, now this is a journey about learning, but a journey about learning what exactly? Well the short answer is this is a presentation about the learning journey of my career through data and analytics as it applies to politics.

And that's the short version of what we're here to talk about today. But you all are here for the long version. You're, you know, in your seats, tucked away for the next 25 minutes. You get to hear the story from the beginning. And it all starts here. Now I am sure you do not have every single electoral map memorized. I'm not expecting that from you. Let me just cut the suspense. This is from the 2000 election, OK?

So what that means is there should probably be an asterisk right down there in Florida. This is the Bush v. Gore election. Now I bring up this election because this is the first time in my life that I remember seeing people really involved and organizing around some big political issue. And I admittedly was a child at the time, but I knew that there was something big afoot.

I was seeing images like this. I would go to the grocery store with my mom. I would see headlines like this. And then ultimately this, and I was like, wait a minute, something big is happening. Something big is mobilizing the country, and I really want to find out more about it. I want to keep pulling the threads and learning more and more.

And so it was after this decisive election that I became that nerdy little kid who loved to watch C-SPAN, particularly the State of the Union on C-SPAN because it didn't have the commentators like the other cable news programs. I could get straight to the point, and I would always try to see how many of the members of Congress I could pick out. I got pretty good at it, you know, just had to brag on myself for a minute.

[LAUGHTER]

But what you quickly learn when you're following politics for a while is that there are subtle changes that happen in these maps over time. You'll see some more blue, a little bit more blue, on the west, pesky Florida, you know, coming around. And we continue to see these small changes, but they actually have big impacts and can determine who the President of the United States is.

And while this is all happening and I'm kind of like watching this from afar, I notice that a lot of headlines are really touting voter data as the thing that saved Barack Obama's campaign. And this is like where it's at. This is how we're going to elect every president from here on out. Sorry about this pop up. Later.

[LAUGHTER]

So I know that this is something that's really big that's happening, right? So voter data, great. Now at this time in my life, I am preparing to graduate which really all that means is I'm trying to find a job. And I was politically active. I told you all I was like the little, you know, political child. And now I'm the political college graduate involved in knocking door to door, organizing, canvassing, all of that.

But through my political connections and the folks that I met, I ended up landing a job as the Data Director for the Indiana Democratic Party. Which I was like, OK, now I'm a big shot. This is super cool. I'm really excited. But there was one major problem, which is that I didn't know how to do any of that job.

[LAUGHTER]

And I mean, I had literally gone from like data entry and now like Data Director. So I didn't really appreciate at the time what a big leap that was. So I had a really serious question I had to ask myself, which was how to learn. OK, so I knew that there were different things I could do, right? So you've got the boot camps. OK, familiar, you know, got it. I could self study. Fair. I could also learn on the job. Excellent.

And I did pursue studying and learning in each and every one of these fields. But you quickly realize there are pros and cons, gives and takes with everything that you do. So boot camp, for example, is really fantastic. Because you get a whole community of people together. A lot of folks are kind of interested in what you're similarly interested in. And it's like a really-- you know, we can have a really fun time meeting people, socializing, great.

But the one major drawback of boot camps I've found is that they are often very intimidating. Sometimes you look to the left, look to the right, everyone seems to know what's happening. And you're sitting there like, yikes. I mean, maybe I missed a big step. I'm not quite sure. But everyone else seems to be getting this, and I'm really struggling. Like why is that? And it can feel very overwhelming.

And then there's self study. And self study is really great, because there are some things where, you know, you go at your own pace. So there are some concepts, I'm the hare. I can zip through that concept. I get this, I get this, I get that. Great. But then there's things where you're really slow. And you might not get it as quickly. But it's OK, because you're moving at your own pace and eventually you'll make it to the finish line.

But self study can be kind of isolating and alone. Sometimes you might maybe cheat yourself out a little bit. You'll speed up maybe through some things that you actually should have paused on. Or maybe you don't know when to ask help or who to even ask for help. So it can be a little isolating in that way.

And then there's on the job experience. Now I personally love this, and especially in the context of politics and doing data for political organizations, a lot of times they can only afford to have one data staffer. And so when you get things right, you are the queen of the castle. OK, you're on top of the world. No one can tell you nothing. Like you are on it. And all your colleagues love you, and it's great.

However, this has a huge negative side, which is that you are extremely vulnerable. You kind of can feel exposed. Now this is especially true within the context of working in politics, specifically around campaigns. Because there is only one election day, the first Tuesday after the first Monday in November.

And if your systems and everything aren't in place, if everything isn't working by that time, you're out of luck. Like if you feel like your data operation isn't tight enough to lead to success on election day, you're left really feeling like, man, have I let my friends down? Have I let my coworkers down in a very intense personal way that you might not feel elsewhere.

So I have all of this experience. I'm, you know, sailing on the winds of self study and boot camp and my on the job experience. And it's tossing and turning me on this wonderful journey. And eventually I arrive at the Island of Lessons Learned. Let me make that a little bit bigger for you.

So what I've come to find is that when I'm learning, there are just basically some major pillars, these three pillars, of community, approachability, and teaching that I try to really keep in mind based on the experiences that I've had building my career in political data.

So the first one is community. Now community is really great, because you get a great network. There are people around you who can keep you motivated. They keep you inspired. You, a lot of times, can rely on these people for jobs maybe to help push you further.

This can be like a list serve that you might be a part of or a particular coding group that you might belong to. It can also provide a sense of representation that might not be found in your on the job experience. And going back to what I was saying earlier about how a lot of times you're the only data staffer, then representation is really huge because you want to find other people who do data work and maybe other people with your similar experience who are working in that field.

So I'm going to tell you a little story about how a community has really impacted my experience and learning journey. So I came into the data world primarily learning SQL. And I use SQL to interact with voter databases, create lists for candidates of the voters that I thought were most likely to support them. And then we would use, you know, paid contact to talk to them, convince them to vote, et cetera.

Now I was using the voter database which already had predictive scores, predictive models appended to it. These models would tell us how likely someone was to vote, how likely someone might have been to support a given candidate. And while I knew the SQL to access the database, I didn't know how to build the model. And I decided, I'm going to build a predictive model. I can be one of these people who's out here building models. I can be one of these data scientist gurus. No problem.

So I ended up kind of poking around a little bit, found a women's coding group in DC, and thought, this is my ticket. Here we go. And I get to the coding group, and this is the first problem that we see. So this is a problem from LeetCode, and it's marked easy. So I'm of course thinking, got it.

So the question here is, given a string, s, and a character, c, return an array of integers representing the shortest distance from the character, c, and the string. And so you can see the output that's desired based on the inputs here. And I'm, like I said, I'm working with a group of other women. Everyone's kind of at a relatively different skill set. But you know, everyone is pretty much congenial trying to work together.

And I'm thinking to myself, OK, well I see this output. So I think we need to index the string. Like I just need to know the index of everything in the string. And then I just would need to know the positions of where all the Es are. I'll do some quick math, subtraction. We'll figure it out. It'll be great, easy peasy.

So someone ends up mentioning this function called enumerate. I never heard of it, so I googled. And I see that OK, I think this will give me what I want. It'll tell me the index position of the-- it'll give me the index for the values in the string. I'll be able to find where the Es are and throw an if statement in there. Great. We'll just, you know, keep moving. We'll do our math at the end and it'll be fine.

So I come up with this beautiful piece of code. And let me tell you, I was impressed with myself when this happened. Because I was just like, I am on a roll. I got it. OK, great, and so I have a list that's going to tell me all the positions of where all the Es are, the word list. So that's just going to tell me where every single letter is.

I printed it out. Boom. I got this output. I was like, this is hot stuff. We are on a roll. Now I just got to figure out some way to subtract those positions from each other, and we'll have the answer.

OK, no we're not. That was actually not at all the right way to solve this problem. I kind of went down a little bit of a rabbit hole trying to figure out how to make this enumerate function work to solve this problem. And I never really got to the right answer. And then the final answer didn't use enumerate at all, and I was just like, well, this was fun.

[LAUGHTER]

But I had a community of people around me. I learned a new function, something that I wouldn't have known otherwise. And I felt really welcomed in a way that made me want to keep coming back. So I'm like, OK, eventually I'm going to build this predictive model and save democracy. It'll be fine. I just need to keep coming back and doing this. And now I have more people to help me, and I'm learning more and more. Great.

But there's also this other aspect of approachability. So when I think about approachability, I really think about centering on joy. Now I say this is especially important in a political context because when you're working on campaigns, it's not necessarily about being the fastest or the best all the time. It's just sometimes really about remembering why you're there, right? Keep calling to mind what motivated you to get involved in campaigns, in elections, and organizing in the first place. So I really try to always remember why-- it's important to always remember why you're there.

I also, when I think about approachability, I want things to feel relevant to my interests. I sometimes will go and I'll see all these examples of data sets that use cars. And they're like, OK, just you know, use this car data set and do whatever. And I'm just like, I don't really care that much about cars. Like they're cool or whatever, but that's just not my cup of tea necessarily.

And then I also like there to be an abundance of low hanging fruit. I want to build momentum, feel like I'm learning and learning and learning. It keeps me engaged, right? I feel like I'm achieving something as I'm moving through the particular learning module.

So this approachability story kind of comes from the similar story where I'm talking about how I'm wanting to build this predictive model. And I'm like, OK, well I've got my old coding group together. Eventually I'll be, you know, running that show and I'll know how to do everything over there. So that's bacon. Let me do some self study, get better at Python, because clearly the enumerate thing wasn't going to help me. I need to learn some more.

And I end up looking into data quests. So I'm doing data quests and I'm going through it, and I'm like, OK they have a good variety of different data sets which I really liked. And because they have this good variety, I'm like, OK, this is interesting. And as I'm going through it, I get to this. Be still my beating heart, like something that is clearly related to politics or government. This was an example that used the White House visitors log as a data set, which I just thought was brilliant. Like finally I feel like I've hit something that speaks to me. I really want to engage with this material.

Now that could have just been a general fluke that made me, that just was that one thing that really connected me there, right? But there were other things that I felt like this program did really well. So as I mentioned, there was an abundance of low hanging fruit kind of principle. So they repeated these simple open steps every time in every module. So I felt really good and comfortable learning this, which was great.

But what I actually appreciated even more was how they listed their instructions and made it so that you could write your output almost to a way that exactly mirrored the instructions. So I could easily identify where I was struggling. And this really made me actually more likely to ask for help I realized, because I knew exactly where I was struggling. So I could feel like I was asking an intelligent question about a particular function that might have been giving me trouble.

And of all of the different pillars of learning, these lessons learned, the teaching one probably surprised me the most. But what I've learned is that when you teach, you're forced to prepare. So you have to actually know in advance of, you know, whenever you're going to instruct other people what you're talking about, which also then requires comprehension.

So you have to understand what you're teaching other people. And you need to understand it probably better than you might if you just did it for your job day to day. When you're teaching other people, you really want to make sure you're fully communicating the meaning of what it is you're trying to get across.

And also I realized that when you teach you become kind of seen as like a leader in the community. And that's a really great way to keep you involved and moving along in a process. Because you kind of become entrenched. Like there are people who you want to stay around for. You might train someone and be like, oh, you know I was thinking about leaving this career path, but now I kind of feel maybe a little bit more invested. Because I see that there's someone that I've trained. And I kind of want to watch them succeed and blossom as well.

So my teaching story kind of goes back to how I learned SQL in order to work in these databases, right? So I had learned it. I worked on the 2016 election. Obviously the outcome was a little brutal. And I remember feeling kind of burnt out and defeated in a lot of ways. And I was like, I don't really know if I want to keep going down this political data path. I'm really not sure if it's entirely right for me.

But I had a mentor, and she encouraged me to teach some SQL basic data skills to a group of activists. So these are folks who had, you know, a lot of experience organizing, but didn't really have much data experience. And we just wanted to expose them to some skills that they might need in order to succeed in the political data world. So SQL is kind of like, you know, foundational to a lot of other jobs and things.

So I ended up agreeing to teach this class of activists. And one of the things that I thought about when I was preparing to teach folks is that there are different ways that people learn. And when we're doing different-- when we're doing a tech training, we're always kind of thinking visually and auditory, right? You hear the words and you hear the instructions that folks are giving. Or you're watching the screen and trying to follow along and all of that.

But I thought, wouldn't it'd be really cool if we could incorporate this kinesthetic, hands on component of learning? Everyone kind of has different things that they gravitate towards, maybe some people would really appreciate a more hands on activity of learning SQL. So I paired up with a friend of mine, Dawn, she's fantastic. And she and I came up together, and we came up with this slide.

And this was just introducing the concept of select and from. So we would present this slide to the trainees. And it's also important to note that we were intentionally laptops down when we did this slide. And all we said was I want to select trainees from this room. And if you were a trainee and you were in the room, you would stand up.

Then we moved on, and I said, I want to select vegetarians from group one. And so we would, making sure that folks kind of understood the concept, if you were in group one and a vegetarian, you stood up. And this really was exciting, because you could see people's kind of faces start to light up. They're kind of understanding some basic syntax and they're moving along.

And so then we moved into this dummy table we created, select alter ego from hero table, and then eventually the column name from table name. Select column name from table name. But this was a really fun experience. Because instead of folks just sitting down there at their computer just typing away, we tried to take something that would have been a little, maybe, easily glossed over in some instances and tried to make it come alive in that way.

And folks really enjoyed it. And so when they saw this example, they knew to expect a list of cities from the data set to appear.

And then finally, I want to talk about Black Girls Code. So I was a Core Team Leader for Black Girls Code in DC. And I initially thought about this slide solely for the teaching aspect of it. Because I think people kind of hear Black Girls Code and think like, oh, well, maybe it's just, you know, teaching teaching. It's kind of more than that.

Black Girls Code is targeted at black girls ages 7 to 17. And what we're hoping to do is get more girls involved in STEM fields. So we want them to grow up and become more involved in STEM fields. And I say that there's obviously teaching in a setup like this. But more critically, every aspect of every side of this triangle is utilized.

So when we first come in to a workshop, we hold workshops periodically throughout the year, and when the girls first come to the workshop, we go over a standard list of community agreements that really reinforce the approachability side of the triangle. So we are talking about things like it's OK to ask any questions. You know, it's be a Carebear. Care for your other girls that are around you, that are learning with you.

And then there's obviously a good sense of community. And we want you to feel very comfortable being there. But I think most critically at the end of each workshop, we encourage the girls to stand up and teach their peers and their parents who would have rejoined the room about what they learned in that given day.

And so we're seeing these girls as young as like 7 or 9 who are standing up and talking about how they learned HTML in a day. And it's really cool and special to see.

So in conclusion, I like the pillars and the sides of this triangle. Because for me what they represent are inclusion and sustainability and learning, which we all, I think, can agree is very, very important. Because taking it back full circle, you know, we know that this is, and data is clearly important in this whole political context. And we know that it's not stopping its importance.

So we need to have a generation of folks who are involved in organizing, who feel like learning is acceptable, is accepting, and inclusive, and something that they want to continue pursuing for them. So that's why I think that this whole triangle is so, so important.

So there are some next steps for me after I depart from here. I really want to learn how to build that predictive model. And it's something that I'm still working on, and I'm so upset that I haven't mastered it just yet. But I'm going to do it. I promise.

And I also want to find ways that we can reduce bias in building predictive models. I've read a lot of different headlines even outside of political context where we're building these models. And they might have different biases, and I want to try and do what I can to help alleviate some of that in the best way that I possibly can with my skills and talents.

And I most importantly want to keep using and refining these lessons learned to foster inclusion in data science overall. I think that this is, what I have here is maybe a good foundation. And as I keep progressing, learning, going along this journey, I'll be able to take more information into this and hopefully maybe refine this model and keep inclusion and diversity in the forethought of my brain.

And some next steps for you all. I challenge you to find two opportunities where you can teach others about something you already know. You might find that you'll end up maybe mastering a skill that you thought that you had. But you're like, maybe I actually don't know that as well as I think I do. Maybe I should find people that I can teach it to. Find people to teach it to as if they've never heard of it.

You can also maybe identify a colleague and encourage them to lead a training session. So how are you uplifting the people that you're managing to become leaders in their own right that will make them want to stay around and stay committed to learning more?

And then maybe you're already a part of a learning community, and maybe you're on that list serve and you never say anything. But you know that you probably should. Contribute! I encourage you, use this as your call to contribute to that community, or join one, or start one yourself.

And so with that, I say thank you. It's been such a pleasure to be here with you all today. And most importantly, I want to thank you for being part of my learning journey.

[APPLAUSE]