Data Science Hangout | Jessie Pluto, Comcast | Taking Initiative with an Idea

Welcome back to the Data Science Hangout. For anybody joining for the first time, I'm Rachel. This is an open space for the whole data science community to connect and chat about data science leadership, questions you're facing and get to learn a little bit about what's going on in the world of data science across different industries too. And so these sessions are recorded and shared to the RStudio YouTube. So you can always go back and rewatch or find helpful resources or share with one of your colleagues as well.

Feel free to always share this invite if there's someone across your team that you think may benefit from joining in the conversations too. We do also have a LinkedIn group if you want to connect with people after the fact. It can be a little bit easier to find each other in there or continue conversations. But together, we're all dedicated to creating this welcoming environment here for everybody. So we love when everyone can participate, and we can hear from you all.

So no matter your level of experience or area of work, there's three ways that you can ask questions. You can jump in by raising your hand on Zoom here. You can put your questions in the Zoom chat. And then third, we also have a Slido link where you can ask questions anonymously. I want to keep adding this. I always frame this as questions. But if you also have something to add to a specific topic, feel free to use those options too, or raise your hand and jump in the conversation as well to share your perspective.

But I am so excited to be joined by my co-host here today, Jessie Pluto, Senior Manager of EBI Strategic Analytics at Comcast. And Jessie, I'd love to have you jump in to introduce yourself and share a little bit about your role and maybe something fun that you like to do outside of work too.

Sure. Welcome, everyone. I'm pretty excited to be here. As Rachel mentioned, I am Jessie Pluto. I'm a Senior Manager here at Comcast. I work in our Enterprise Business Intelligence Department in Strategic Analytics. And my focus is really on video and content. And a big part of my job, I've been here for almost five years. And my goal every year has remained pretty consistent. Every year, I am really trying to save on the programming costs. Programming cost is the largest cost that Comcast has. It's outpaced even labor. And so really, I am trying to save, you know, a couple hundred million dollars here, a couple hundred million dollars there, you know, no big deal, get to those B's every once in a while.

I actually have a background. I studied psychology in undergrad, which is why I think I really like the behavioral part of this job, looking at, you know, customer behavior, trying to figure out like, what programs do people really care about? I kind of studied psych in undergrad, I did a master's certificate in applied statistics afterwards, got my foot in the door in a GIS job, trying to do some like market research around like, where can companies, franchise restaurants really open up a new location that would impact sales to the existing markets. Worked for a brief period at a pharmaceutical research company before I came here to Comcast for the last four and a half, five years.

That's awesome. So casual, save $100 million here and there.

Here and there, you know, no big deal.

Pretty basic stuff. I like to go hiking, I love to get outside. I live here in Center City, Philadelphia. So I do love to just get outside the city, get to a national park, you know, lose that cell service, lose the Wi-Fi and just go hiking for a couple days or a week here and there.

Transitioning into people management

So yeah, I just love that idea, by the way, about creating, like being intentional about creating a sense of community, like, how do you do that? I really feel like I kind of like took a lot of that where it's okay, like, what are the, you know, little things that I can do to try to, you know, be intentional about like, what is the culture that I want to create here for my team.

I do have the advantage of coming from a, I had a manager before I got, you know, went into people management, who was just like an exceptional manager in terms of creating a culture that I really wanted to kind of almost replicate. And it's a lot easy to kind of like borrow than it is to create something from scratch. And so I really borrowed a lot from some of my previous managers. One of my big, you know, principles, again, that I have borrowed from previous managers is this idea of one recognizing false deadlines, and not necessarily trying to hit these false deadlines, if it's not necessarily something that's going to make a business impact.

And then just really like borrow often, share freely, you know, take code that's out there, share code with other people see what how somebody else has done something in the past and really try to kind of leverage what's already been done out there. I think that's been really helpful for, you know, just kind of like building like a collaborative culture where it's like, okay, we may have done something for the first time, but we borrowed this part of it. So like, let's give a shout out to this person. You know, let's put our code on GitHub, let's go around to other team members and just kind of share what we did, why we did it, what kind of we learned throughout the process. I don't, I might own this project, but, you know, there's no reason to be possessive of knowledge. Let's share that as often as we can.

I don't, I might own this project, but, you know, there's no reason to be possessive of knowledge. Let's share that as often as we can.

And that's something that I borrowed too from the RStudio community, just growing up in that community. Well, not growing up, but you know, learning, getting into analytics in that community has been just such a great learning experience for what works. And then really, it's just kind of like doing, I think, like you said, like those touch-ins and trying to make sure that I have frequent, you know, check-ins with my team.

It can be very, very difficult, honestly, especially when you get a new team and you're trying to transition out of doing the hands-on work yourself. This is one of the things I struggle with the most is making time for, you know, the people on my team who need feedback, who need to be moving forward to be successful. And sometimes that is at the expense of, okay, I can't do my own code that needs to get done, but this is the right thing to invest my time in right now is making sure that they can move forward and that they have what they need.

Staying available for your team

So I do have scheduled one-on-ones every week on Monday with my team, you know, both down and upwards just so that we can like make sure that we hit the ground running on Monday. And then throughout the week, we have a couple of other team meetings, you know, like where we meet as a group. I think during COVID, one of the things that we got used to, you know, working from home is just the idea of, okay, we don't necessarily need to have time on our calendar. We need some sort of daily check-in where it's like, okay, do you have what you need?

Part of it is honestly just making sure that I am available, that I have time on my calendar, that I'm not completely overbooked to the point where if my team needs some feedback that I am not available for two days, that's not a good way to work. And so some of that is honestly just blogging off some time on my calendar and making sure that I have time every day that I can be available.

I do like mentorship program for like Big Brothers Big Sisters as well. And one of the things that I learned through that is just how important it is when you first are matched with somebody in a mentorship role to be available, to just show up. And I think that especially when you get new teammates, new people reporting to you specifically, like it's so important in the beginning to just show up to just be available.

Prioritizing mentoring over meetings

I'm in a very, very fortunate role where I have a small team and we make huge business impact, financial impact on the company's bottom line. The impact that my team has, the impact of what certain projects, how certain projects can benefit the company, I think does give me a little bit more power than most people in a typical senior manager role with a new team at a large corporation like this. I can push back and say, listen, this project, this is a, you know, $250 million annual contract that we're working on right now. This is our main priority and this is why it's important.

I do have to be proactive. I can't just start canceling meetings that are on my calendar every week. I have to be proactive about, okay, here are the days and times that I am trying to be consistently available, trying my best. And it's not easy, but trying my best to manage across many, many stakeholders.

Tools and calendar management

Yeah, I actually, I've tried so many tools and the thing that I recently have kind of circled back to is just trying to keep it simple, stupid, you know what I mean? So like it is mostly Outlook. There was a plugin. I hope we don't have anyone from Microsoft that I'll offend here, but I tried some Microsoft plugins. I think there was something called Viva that would like block focus time on my calendar. And it was constantly getting like overlapped with other meetings and just like not a convenience, like 15 minutes here, 20 minutes there. And like, that's not a good way, especially coding. That's not a good way to work.

So I ended up just kind of literally, I planned it like a month ahead of time where I, okay, for this month, it is what it is. I'll manage them how I can. I'm going to reply tentative to the meetings that I don't have to be at. Unless somebody is in the invite saying, this is what I need you here in this meeting to do, I'll reply tentative. But it was like literally a month ahead of time. Okay, the Viva thing wasn't working. I'm going to block a big chunk on Mondays, Wednesdays, and Fridays of like heads down. This is the time that I'm available.

My employees know that that time I am available and everybody else on my Outlook calendar sees me as busy. And then I make sure that I leave open time. So people do have time to schedule meetings. A big part of it is really just learning how to say no politely. Just kind of like setting an expectation that what can I help with? And do you need me there? And maybe I can send somebody else, or maybe I don't need to be on this one, but I need to be on the next one.

And on the tooling front, it's really just Outlook at this point. I use JIRA mostly for project management, but it's really mostly just Outlook and Teams. So Teams chat, Teams meetings.

The breadth of the team's work

I think I mentioned a little bit, but our big priority is really kind of predicting, okay, we have these contracts that are expiring in the next 12 to 15 months. It's coming up with some, you know, we have like an analytic suite that, you know, me and my team have developed, which basically predicts how many customers would, you know, churn, drop video if we were to lose specific programming. It does a lot of additional like analytics in terms of just descriptives around what are the popular programs, what is the seasonality, the tonnage, like how much do people consume this, but then going beyond tonnage, like what's people actually care about.

We really see that, you know, people's passion is with, you know, are you tuning live to like House of the Dragon? You probably care a ton if you're watching that literally at like 9 p.m sharp compared to, okay, you just binged, you know, the last, I don't know, eight season of Friends or something. You know what I mean? Like it's just different types of, the total hours might be the same, but it's not all equal.

A lot of our other, you know, operational responsibilities line, okay, we have this contract that's expiring. And even if we're not planning to drop the content, we have to be prepared legally for if that situation were to happen where we don't come to terms. And so we do have to sometimes fill in a lot of operational roles in terms of preparing, making sure that the call centers know how many calls they should expect. If we are to go dark on some programming, make sure that they have all of the customers, you know, in the system tagged as like, these are the heavy viewers of this. These are the passionate folks that you should be aware of.

And we've really, in my team invested a lot in what can we automate as much as we can, you know, take tasks that we do multiple times and just make it reproducible and as light touch as possible. We've kind of invested in, in making sure that we can do that easily, efficiently, and, you know, with as little time effort as possible.

Taking initiative with a new idea

It is a constant battle. And I feel like it is the place where like it requires some strategy because you can't do it all the time. You have to like really, I think, recognize when is the right time to pounce on an idea. You have to have the ideas. Sometimes you have to sit on them for a little while and then you have to just be able to like recognize, I think, that opportunity. And sometimes it's okay to make mistakes. Like sometimes you push something too soon and, you know, it might hurt a little bit, but don't let that kill your motivation.

I hate PowerPoint. Like I hate copy and paste so much. And for a long time, you know, some of the reports that I was doing, like just to give you an example, Paramount is like 30 individual networks. And so sometimes we're repeating views for, you know, 30 networks times like five different views or something as a bare minimum, too many slides, too much copy and paste. And I had made a, you know, push for, oh, we should convert this to markdown or Shiny or something. And same thing. It's like, well, our stakeholders don't know how to use that.

And COVID hits and sports viewing kind of just went away. There's like COVID hit, pandemic hit. And we had so many questions coming in about what do we do? What's going on? What are customers doing? What are they watching? You know, all sports just kind of went, you know, completely off the air immediately. And I realized like, this is my moment. I created a, you know, this is back, we were in Hadoop at the time, we've since moved, but we were in Hadoop at the time. I set up some like shell scripts, like fast scripts to, you know, run at certain days and like scheduled it and like created this like pipeline that I had been messing around with for a while, created an automated R markdown report.

And basically I was able to give the information that the business was asking for very, very quickly. And they learned how to use it because it's like, they really wanted it. And it would have been too much for my very small team to be reporting every single week on like what is happening this week. And so basically kind of investing. And I will admit that I spent a little bit of night, you know, time on this project. But it was a passion project where I was like, all right, this is my moment to prove that like, this is a feasible solution.

I was able to show that, okay, this is done. We've automated, we can use this for other things, for other questions that business owners want every single week, for data that they want every single week. And still have time left over for those ad hoc questions, those one-off things that are really important. And so I think there's just this kind of strategy that you have to really kind of always be thinking about of when is the right moment to, you know, maybe not necessarily ask for approval, to just do something and show the value.

And so I think there's just this kind of strategy that you have to really kind of always be thinking about of when is the right moment to, you know, maybe not necessarily ask for approval, to just do something and show the value.

And hopefully you have leaders who will support you in those decisions, if it's the right one in the right moments. Because that's really, I think that's what I've been able to, those, I think like little windows that I found to be able to kind of push some of the different tools that, you know, were not really, I think, supportive in the beginning of my career here.

QA and team structure

So my team doesn't actually, we do in enterprise business intelligence have a reporting team, um, because the viewing data is sensitive. And we kind of like, again, I think I said like my team, whereas a lot of different hats, like sometimes we get into the engineering space. Sometimes we get into the data quality space. A lot of times we do even get into the reporting space. The COVID report, that was the solution that fit the question at the time. We didn't continue to support that. If that were to continue to be a, you know, solution that needed to be out there every week for our business owners, it would actually transition to another team in EBI, the reporting team.

Strategic analytics really is more focused on, you know, these are the questions that the presidents and VPs of the company are asking. How can we answer them? And then once it becomes more of a, you know, even like a model, for example, we might build a segmentation and strategic analytics, but when it's time for deployments, you know, we might hand that over to the data science team to actually, you know, get into a good deployment.

Building a culture of learning

Yeah, I actually, I've tried so many tools and the thing that I recently have kind of circled back to is just trying to keep it simple, stupid, you know what I mean?

Yeah. So at literally as of yesterday, I finally got the green light from our COO to start a data scientist and analyst group. I'm not an official manager of these people, but there's a lot of disparate systems and, um, workflows in the organization. We don't have a common like data warehouse or single source of truth. Um, we're a family of companies. There's seven companies and we're all over the place. So part of this group is a, to figure out what people are doing. Um, but also much to Brittany's point, like how do you get them to step back from the hustle and bustle of checking boxes, being more of like a compliance minded.

And, um, we're going to have, um, these groups are going to be broken up into what I call show and tell modules. We'll build, somebody will literally just be able to come up and say, this is what I've done in Power BI. It doesn't have to be anything fancy. We just need to know for transparency sake, what people are doing. And then, um, we'll also have learning sessions where somebody who feels like they have at least a periphery knowledge of a software or a tool or a programming language can show other people what it what's capable of and what they've utilized it for.

I think right now we're shooting for every other week, Brittany. So, you know, one hour every other week, if that's not doable, then there's something very, very wrong with their management. And I'm also willing to go to bat and start, you know, duking it out with managers if I need to in a polite and political way, but that that there's no reason for anybody at any job to be, you know, eight hours a day nose to the, to the, um, or pedal to the metal, just inundated with work. If you don't have room for improvement and or skill management or skill improvement, then you're going to be stuck in this rut. And that's when unhappy employees leave.

Workflow and version control

Yeah, I love talking about workflow. So, um, just from a, like, where does work come from standpoint, like going to like the very, very, very beginning, um, we actually don't have a JIRA intake for it. So it's not like anybody can just submit a request and then all requests are treated equally. We, you know, strategic analytics, like sometimes I described it like the internal consultants of Comcast, you know, like we are in the meetings with, you know, the president of content acquisition. And so we are able to kind of hear, what are the questions that are being repeated over and over again? Where's the value that's kind of like on my team to figure out what is the biggest value that we can add right now?

And then basically my team, you know, we have most of the viewing data is in, we just moved recently to AWS. Our data is an S3 storage. And, you know, we do a lot of engineering to kind of set up, you know, jobs that run, you know, every month or so. And basically we'll do a bunch of aggregation in, you know, Databricks, for example, and a Databricks notebook with Scala, um, or, you know, Spark or something like that.

My team has recently started working. I think I mentioned before that I kind of was doing this mostly independently for a while. And so when I started working as a team, I was really excited to finally have the chance, like we're going to work from a repo. Like everything is in the GitHub repo. We're all like working, you know, we started off just on like GitHub desktop, honestly. And, you know, just working from the same repository altogether at once. It's just like such a fun experience. And of course I messed up one of my, I messed up the first one, the first process, of course, which I actually felt really good that I was like, okay, both of my, you know, new teammates did awesome. I screwed it up and I got to kind of be the person who, you know, made an example of what not to do.

That's kind of my new, I think passion is just like making sure that we all have like a good solid repo that we're using proper version control and, you know, contributing in that way. Cause I just think it's such a better workflow, honestly, than what I had been doing before, even though I was using packages and trying to kind of like standardize and make all of my code reproducible. It's different when you're doing it on a team. And I will say like, if you haven't had that experience, it's definitely, you know, uh, scary, but in a very good way.

Sharing failures and psychological safety

With a big smile, usually laughing. Of course it was me. I think when I was earlier in my career, I was always so nervous of making a mistake, especially publicly, or God forbid, like as a manager with people who reported to me, it seemed like such a scary thing. And I don't know if it's just like years on the job, if it's just like the results of doing it many times. I don't really have a good sense of like, why I've kind of grown out of that, but I have grown out of that to the point where it's, you know, I guess maybe it's just like having enough successes under your belt and having a reputation that you can, you know, stand firm on.

But I definitely feel a lot of, um, psychological safety, I guess, to kind of fail, fail publicly. And, you know, hopefully nothing really, really bad has happened. Luckily I don't work in medicine or anything where like lives are on the line. Mistakes happen. I think that it's a good learning opportunity. As long as it's not everything that you do and you have a reputation as somebody who can deliver and, you know, isn't constantly making big mistakes. I don't think you need to worry so much about it.

Hiring and community involvement

Yeah, I'm still learning a lot of those lessons. I think I have been doing some hiring and interviewing and it can be very difficult, I think to, you know, in such a quick, you know, 30 minute interview or so, like really get to know a person. Um, not just like in terms of their technical skills, but, um, it is really, really, really hard. I think to like, look at a resume and to necessarily, you know, get a good sense of like, is this the right fit?

I think whenever possible, you know, just trying to be more involved with the community and have that kind of, you know, if you're not involved with R-Ladies Philly yet, I'll give a shout to R-Ladies Philly. That's a great place to get involved. Even like the RStudio conference. There's like lots of ways where I think I've been able to benefit from just already being in a community. Just being involved like that is, you know, I don't know if it's going to work yet, but that's my strategy right now is just kind of like some of the other things at my, you know, in my career, just trying to take ownership of, okay, I'm not going to leave it all up to the process that exists in a large company.

How the team was formed

So good question. Um, strategic analytics. So enterprise business intelligence at Comcast, I think is like seven or eight years old. It's not that old. It's relatively new. It started with, I think like 12 people or something. Um, we grew really, really fast. And one of the big wins, this is before I was here just to, just to call that out. But before I was here, one of the big early wins that we had was actually this kind of work that I'm doing now. It's getting access to the viewing data. It was a big win getting involved with the president of content acquisition and building that relationship with a huge win where we were able to very early in my department's, um, history, like show a huge value to the organization.

Um, but in general, what happened was basically I was doing this content acquisition support, as well as a bunch of other video and content support. And my manager was a very, very good mentor and him, it was a team of three, him and the other woman that I started working with on the team. They went and started their own team in the same department in strategic analytics, where they are completely focused on training new hires, on giving them, uh, you know, cross training across all the different groups in strategic analytics. And so basically that kind of led me into this role where, okay, now I am leading the video and content team and doing all of that support and kind of starting my own team from scratch.

I have two people right now and I am looking for one more.

Cross-team collaboration and documentation

Yeah, we do do projects together pretty frequently. A lot of times we'll team up. So we have like a finance team in my department. And so like a lot of times we'll team up with the finance team. We actually just created a new team called channel analytics coming out of the collaboration between the product side and the finance side. So we do a lot of, you know, whenever we can collaboration. Sometimes it's within strategic analytics. So like, you know, we'll do a lot of, you know, product finance customer experience. All of those things. There's a lot of collaboration that happens there, especially even code sharing.

Yeah, my team, I'm fortunate my team is small. And because we don't have too much ownership of like the actual data, you know, data products, like the viewing data necessarily, we don't have to stress documentation too much. I do try to just like, be better, especially working as a team. I'll be honest, I was terrible when I was working mostly independently. I've been trying to be a lot better now, just like having a team working as a team, commenting code. You know, pushing updates to GitHub, every time I make any updates, things like that.

In terms of like actual documentation of workflows, we don't do too much of that on my team. I think that's a, it's a huge amount of overhead. And we're more focused really on being fast and effective. And when there's larger products, larger projects, over a longer period of time, honestly, just making sure that it's reproducible, less so than documented.

Yeah, one thing I guess we do have a, you know, a blog post that we have internally that is, you know, hosted through I think GitHub pages, but we do, you know, if we figure out like a new way to move. One of the things we had to figure out when we moved to Databricks was, okay, how do we move data from, you know, cloud location to an on-prem location. And so we figured out how to do that. And then, you know, just kind of like writing a script that's well commented, posting it on the blog, sharing it around with the team. So like things that do impact, you know, more throughout the organization, we do try to kind of put some documentation together for that and post it on the blog post right now is kind of the place where we're putting it.

We do some stuff locally at the last step for some reports where we're pulling in data from the cloud. It's already aggregated. Most of the processing happens in the cloud, but sometimes we do use our local R every once in a while.

Well, thank you so much, Jesse. It's been really fun getting to learn about the ways Comcast is using data and also how you're building that culture and showing value on the team. If people wanted to connect with you, LinkedIn, R-Ladies Philly, Twitter. All of those are good.

Okay, awesome. Well, thank you so much. Really appreciate it. And thank you everybody for all the great questions, too. I feel like there was so much energy here today. Really, really appreciate you all being here. We'll be back next week. We'll be joined by Marcos Huerta, manager of data science at CarMax as well. So hope to see you then. Bye, everybody.