Sheikh Shuvo (00:02.774)
Megan, the very first question I have for you is that you described yourself as a data journalist, which is a term I just absolutely love. I feel it's so descriptive. Could you share a bit about what that means to you and how it relates to your world?

Megan (00:20.157)
Yeah, definitely. Essentially what it looks like for me right now is I create content about all things data. So I work for Alteryx, which is a data analytics software company. And I'm using their products, doing use cases. I kind of have my own brand on LinkedIn and creating data content for that. A lot of people will come up to me

think that I have a background in journalism. I don't, so in some ways, you know, I don't wanna say a fake journalist, but I'm learning as I go. I started out writing a blog and that took me here. So it's exciting and I get to also host a podcast for Alltricks, so written content, audio content, all kinds of data content.

Sheikh Shuvo (00:53.59)
Hahaha.

Sheikh Shuvo (01:16.09)
Looking at your content, there's nothing fake about there. I think you can count yourself as a bona fide journalist slash investigator. Now, you started your career in industrial engineering then shifted to data analysis. What were some of the inflection points along the way that got you where you are now?

Megan (01:23.477)
Thanks.

Megan (01:38.705)
Yeah, so I really loved industrial engineering in college because it was a mix of a lot of things. We got to take courses in a lot of engineering disciplines as well as supply chain and database management. So I off the bat had a lot of interests. So I went into consulting, I guess before that, I'll backtrack to my internships.

They were both manufacturing focused, but during those internships, I realized the projects that I liked the most was when I got to work with data. So whether that was diving into a huge Excel spreadsheet or, uh, learning SQL and starting to understand their data warehouse, I really enjoyed those projects specifically. So then fast forward, I accepted a job in technical consulting, uh, for an accounting firm, and that wasn't quite.

a great fit for me. It wasn't quite what I was expecting. I think I like to have a role that's a little more behind the scenes. So after I left that job, I did a data science bootcamp course and online during the pandemic. So that was a big inflection point. I really enjoyed that course. And throughout that course, I was also starting to write some more articles, kind of write about what I was learning as I was learning it as a way to

cement that knowledge and I fell back in love with writing. And then I started a job as a data analyst at Stanley Black and Decker. So that was another inflection point and kept up the writing and eventually got, Alteryx reached out about this data journalism job. And I thought, well, that's kind of like what I like to do for fun, but as a full-time job. So that sounds awesome. So been a bit of a winding road, but.

It's been fun.

Sheikh Shuvo (03:36.31)
best roads are now when you were deciding to take that take that bootcamp class. How certain were you that hey, this is the thing for me. What were the other options you were considering at that point?

Megan (03:54.233)
Yeah, I think I was looking at job descriptions of data analysts or data focus roles. I knew I kind of wanted to work with data just based on the projects that I did like, and even in consulting had some opportunities to work with data. And so that was kind of my focus. And then in those descriptions, I just felt like I wanted a little bit more background, understand Python and SQL and the coding languages more and some of the

theory behind all of that. I also enjoyed statistics in college, one of the few. Everybody else complained about it, but I liked it. So that was like further confirmation. It was like, oh, this is based on statistics. You know, I think I'm gonna like it.

Sheikh Shuvo (04:36.48)
Yeah.

Sheikh Shuvo (04:40.042)
Now, when you were picking the particular boot camp that you chose, there are lots of different options there. What were some of the criteria you used to pick that particular boot camp?

Megan (04:53.793)
I mean, cost was a big thing. I ended up picking one called Thinkful, which I think has changed names since then, but it was lower cost. I did their more flexible program. I can't remember the name of it, but it was lower cost and you got a mentor assigned to you. So that was like really helpful to have him to learn from and...

connect with while the rest of it was pretty self-paced and everything. Um, so yeah, and I, I liked the content that they, they put out and they're very focused towards, we'll help you get a job in data. So that was a selling point.

Sheikh Shuvo (05:38.007)
Nice.

Awesome. Now, since you initially started that journey deeper and deeper into the data world, have your preferred tools for data analysis changed at all? Where did you start and what does your stack look like now?

Megan (06:01.137)
Yeah, I mean, when I was going through the course, I was learning Python and building models with that. And for data science at least, that's what it looks like now is that I use our company's AutoML tools. And they're great and they're fun to work with and it is helpful to know.

more about modeling so you can go in and, you know, customize it. But it's also exciting that people that are more from the analyst background can start modeling, um, easier. So I guess, yeah, on the data science front, that's how it's changed in terms of analytics, you know, gone from learning Excel to using more SQL and Ultricks designer. And so, um, I know like.

as an analyst learning SQL is important to kind of optimize the queries before I got the data into kind of the ETL process. So yeah, I think those are my two kind of progressions of the stack I use.

Sheikh Shuvo (07:11.879)
Yes. You mentioned AutoML. Are there any broader ways that you incorporate AI into your workflows?

Megan (07:23.493)
Yeah, there's, we also have this other product. I'm not really trying to sell here. I just, I legitimately do use our products in my workflow. Uh, but, but yeah, we have auto insights, which is basically automatic dashboarding and then it uses ML behind the scenes to pull out. It'll, you know, write full sentences and say, I basically put our blog data in there and it says in the past week.

Sheikh Shuvo (07:31.286)
No, of course. Yeah.

Megan (07:53.573)
blog traffic has gone up 9%. Here's some likely causes of this. And it highlights like these drivers of change. And so there's like machine learning working in the background there. So that would be the biggest thing that I use in my workflow now for kind of content strategy and analysis. And I've used the auto ML tool a little bit more for like, writing articles about it. But

It's been fun and I can kind of like pull some classic machine learning data sets examples that I've maybe have touched in Python and see how AutoML their tool works. And, um, yeah.

Sheikh Shuvo (08:37.954)
Nice, awesome. Yeah. It's like having four hands then. Going back to Alturex overall, clearly a very well-known data science platform that's been around a while. And one of the taglines of the company is to enable reliable analytics everyone can trust. Could you speak a bit to what's meant by trust in data?

Megan (08:41.145)
Yeah, it is.

Sheikh Shuvo (09:06.714)
In what cases would data be unreliable when you're working within the company?

Megan (09:13.545)
Sure, yeah, that's a good question. There's lots of examples I could pull from. I think when it comes to modeling, it can be hard to trust a black box model that you don't know what's going on behind the scenes, or even an automation that's totally based in code for data ETL.

It can be hard for non-technical people to understand. And so I think a piece of it is that Alteryx has like a more visual interface where you can see step-by-step the process flow of where the data is moving and check at each step. It's just a little bit easier to grasp and explain and build trust when you can say here's each processing step we're putting the data through.

Um, I think another example is we see people all the time who do. Report like important reporting processes and Excel, like important financial stuff. And it's really easy if you have this complex Excel workbook to make a mistake. Um, and to, um, you know, create.

a situation where there's like the governance piece just isn't there. So I think that's another thing that we're seeing people more interested in, in governance and how do I make sure my processes are well governed and I can trust the data results from that. So I think that's another example of getting it from Excel into a more standard, reliable Alteryx workflow that we can automate and

document and they're coming out with tools for auto documentation, using generative AI to describe what's happening in the workflow, which I think is a cool application as well, a cool thing to continue to build trust with stakeholders that might not be touching data every day.

Sheikh Shuvo (11:26.282)
Awesome. Now, looking more broadly then, one of the other things you do at your day job is posting the Alter Everything podcast. I've listened to a couple episodes, really loved the guests that you have there. After posting a couple dozen of these, has there been anything that's surprised you after chatting with so many different

Megan (11:43.189)
Thanks.

Sheikh Shuvo (11:54.646)
practitioners on how they live their data lives.

Megan (11:59.805)
Yeah, definitely. We do get a variety of guests in different types of roles, different industries. But something that has surprised me is how much it comes up about all kinds of data projects. In my mind, every data project should start from a clear business problem. And I think it's surprising how it comes up a lot across industries that.

companies kind of start with a solution and all of a sudden you lose sight of what was the business problem here. It gets really complicated. And especially with all of the hype around AI, it seems like sometimes companies are starting with, well, we want AI and then like working backwards to find the business problems. Whereas like you need the business problem driving it.

Sheikh Shuvo (12:55.572)
Yeah.

It's like 90% of startups right there.

Megan (13:02.433)
Yeah, you need the business problem driving it and you can end up in a spot otherwise where you have this dashboard you made or this model you made that people aren't using because it doesn't solve a real clear problem. So that's been one thing. I think another thing that stands out to me with our guests is just the diversity of backgrounds that people have working in data. I think it's cool.

me coming from an engineering background, taking this kind of winding path. It seems like there are a ton of other people that have had winding paths that lead them to data. I mean, I haven't really heard of many colleges that have a data analytics major per se, and I feel like data can kind of find you and your role. If, um, maybe you're in supply chain, maybe you're in finance, maybe you're in HR and all of a sudden like.

you don't have a resource and you have to start analyzing the data. And I feel like I ended up talking to these people who got presented these data problems and were willing to learn and upskill and have worked their way into this current role. So I feel like I see that a lot as well, especially with like analytics, specifically having that combination of the domain knowledge of where they were before, but now they were willing to.

try out this new tool, learn this coding language, and all of a sudden they're leading analytics for their company now. Maybe not leading, but you know, it's crazy where people can start out, so that's another thing I've noticed.

Sheikh Shuvo (14:39.276)
Yeah.

Sheikh Shuvo (14:43.65)
Well, looking at your work on just being a data journalist then, are there any business problems you've come across that you'd love to apply data science lens to? What have you solved about your own job?

Megan (15:01.457)
Um, yeah, that's a good question. I think.

I mean, using the auto insights tool that we have for the content strategy. I mean, part of the business challenge there was just that from our Altrix community, there's tons of data. And I feel like a lot of companies are in this situation where now they have tons of data so much, they don't know what to do with. And so, like using.

tools in a smart way, using tools that have machine learning built into them to even pull out here's something that could be insightful. You know, you don't have the time to combed through the data to look at every exception or whatnot like a data analyst might have the time for. But so how can you leverage AI to show you what could be important and then you can start from there as opposed to starting.

You know, with a huge CSV file, not knowing, you know, what, what to do next.

Sheikh Shuvo (16:12.815)
Awesome. Looking at your personal work then with all the great articles you post on LinkedIn and Medium there, is there a process you have for finding inspiration? I'm wondering where the ideas for your content come from.

Megan (16:33.061)
Yeah, it's definitely from a lot of places. It definitely started out when I was taking the course, I got a lot of inspiration from the course topics and also from talking about those topics with others in the course. And then got more into medium and the towards data science publication. Think reading what other practitioners and other experts are doing is always super helpful.

and insightful, so keeping up with, you know, the top publications, also reading in general just helps you be a better writer. So kind of a two birds, one stone situation there. And I mean, on top of that, like consuming content on LinkedIn and really trying to learn and then combine that learning with

what I'm seeing in my day-to-day job and saying, what problems do I see? Where do, I think I did an article on how to maximize your efficiency as a data analyst because I realized I had been wasting a lot of time or whatever and there's things that I wish I would have known. And so kind of decided to write about that. And so I think.

the inspiration can come from a lot of places and I'll just jot things down sometimes. I probably have like 40 unstarted articles, like just titles of articles on Medium because I'm like, oh, this sounds interesting. And then, you know, maybe I don't know everything to write about it now, but I could come back to it. Or someone might say something at work that sparks that. Or someone asks me a question at work, I answer. And I think that could make a good post. So it's, I mean, it's a lot of things and...

Sheikh Shuvo (18:08.112)
Yeah.

Sheikh Shuvo (18:24.552)
Awesome.

Megan (18:28.261)
I think some people, at least on LinkedIn, have this formula of I sit down for two hours, I write my five posts, I target one to customers, I target one. I don't quite have that formula. It's a little more piecemeal together and it's like whatever is exciting. If I read an article and I'm excited about it and I have something meaningful to add, I might write a LinkedIn post. If there's a lot of experiences I have, I might do a blog.

Yeah, I don't know if that answers your question.

Sheikh Shuvo (18:59.63)
No, it does, it does. It's a lot more organic and way more fun this way too. Well, looking at the ideas that you've jotted down recently, what are some of the areas of data science that you are focused on for right now and doing more research in?

Megan (19:04.506)
Yeah, it is way more fun.

Megan (19:21.797)
Um, well, I just took this course, uh, on Coursera about data ethics, data science ethics. And so that really was inspiring and really interesting and had some, some topics come out of that, that I definitely want to learn more about, dive into, write about one of them just being like bias and algorithms and how, um,

Megan (19:50.777)
Oh, sorry, I'm blanking on the word for it. Basically how bias can get kind of ingrained and encoded into algorithms where it's like a self-fulfilling prophecy situation, stereotypes, things like that. So I thought that was really interesting. There was also some course content about when quantification kind of loses its meaning.

Or like when you need more context, when you can be misleading with data. I think that's always like an interesting topic as well. That I want to explore some more.

Sheikh Shuvo (20:30.698)
Yeah, that's awesome. On the last topic, my favorite book still remains How to Lie with Statistics. And so there's definitely lots of ways you can take things in.

Megan (20:39.893)
Mm.

Megan (20:44.327)
Yeah, definitely.

Sheikh Shuvo (20:45.166)
Well, yeah, well, the very last question I have for you, Megan, is looking back on your earliest times when you started getting into the world of data, what were some of the maybe mistakes that you made all along the way that you'd advise other people early in their career to watch out for?

Megan (21:10.437)
Yeah, I think for me, I think it's my personality, but I wish that I wouldn't have gotten as stressed out as I did. I think if I could go back and talk to myself when I first started as a data analyst, I would say, you know, you can do the best you can do, but sometimes there are bigger data issues that you can't solve or that are not your fault that come from.

being a huge organization that has acquired things and there's all these mixed systems or there were decisions made about IT 10 years ago that make your job hard and like, make it hard to deliver a dashboard that people are gonna trust because of, yeah, because of the whole system that you're in. So like, if I could go back, I would probably say, you know, start.

Sheikh Shuvo (21:55.382)
Yeah, that's great advice.

Megan (22:06.665)
building contacts in IT and data engineering to try to help with what you can for like your certain projects, but don't take on the weight of the whole company's data structure.

Sheikh Shuvo (22:23.353)
Yeah, that sounds demoralizing if you do.

Megan (22:27.145)
Yeah, yeah, definitely.

Sheikh Shuvo (22:32.274)
Well, Megan, this has been super fun. Thanks for taking the time to share about your experiences and your world.

Megan (22:39.901)
Yeah, thanks for having me, Sheikh.

Please note that the following transcript was generated by an AI tool and may contain errors, mishears, and misspellings. While efforts have been made to clean up these inaccuracies, the transcript may not perfectly capture the spoken content. For the most accurate representation of the conversation, it is recommended to listen to the actual podcast episode.