In this episode Ray Givler, Senior Decision Consultant for Highmark, talks about creating data visualizations for a healthcare organization using Tableau. Ray talks about how he finds fulfillment in his career through Autonomy, Mastery, and Purpose. Listen to the episode to learn Rays advice for new data scientists and how to start your career in data visualization.
Episode 16 Transcript
(introduction music) - This is Buffalo State Data Talk, the podcast where we introduce you to how data is used and explore
careers that involve data. - Hello, and welcome back
to another episode of
Buffalo State Data Talk. I am your host, Brian Barrey, and thank you for joining us today. Today, we'll be talking to Ray Givler, a senior decision support
consultant at Highmark,
a healthcare company. Thanks for joining us today, Ray. - Sure, sure. Glad to be here. - Awesome. Alright, so,
could you start us off by giving us a general
overview of the work you do
as a lead senior decision
support consultant? - Sure, sure. I do a lot of data
visualization in Tableau. That's my main duty and
that's kind of end-to-end on that from eliciting requirements
and then iterative development,
data prep, validation, pushing that final dashboard to server, and then ongoing maintenance
of those same dashboards. I also mentor junior
developers on my team, as well as some others for the
internal Tableau user group.
I do demos and tips and
tricks occasionally on that. - Now that we have an idea of what you do, can you tell me about what a typical day or week looks like for you? - Well, my schedule varies a lot.
The beginning and end
of the week are freer for me to actually get stuff done, meaning you know head down in
Tableau and building vises, and the middle of the week
tend to have more meetings. - Cool. So you told us that
you work with a team of people.
How would you say your time is split between working with your team
and working independently? - Well, outside of meetings,
there's very little work we do on the development
side as group work. We support each other when we're stuck
and we kind of work in
separate dashboards there. So each person's kind
of doing their own thing on the BI side. So there's not a lot of like two people working on the same dashboard
and then integrating that together. Although occasionally that happens, that's not the usual. Our biggest integration is with the data engineering
sub-team under my manager.
So they're doing an awesome job of pulling together and data internally from multiple sources and
creating reporting tables that are much easier for
us to digest, you know, BI friendly data sources.
It makes a big difference. The meetings that we typically have during the week can be about
a lot of different stuff, primarily data issues and requirements, maybe things like customer feedback
on a new dashboard iteration. We meet for customer training. So we're big into doing
whatever it can take or whatever it takes for user adoption. So those internal training
type things are some
of our meetings and we
have ongoing office hours, we call them, for any
questions that come up. Sometimes there's ad hoc meetings for problem solving and dashboards, like just ping somebody in Teams and
'Hey, can you look at this?
I can't figure it out'. Or sometimes it's not even just a problem, somebody gets excited about
something they've done and I'm like, 'Hey, wow,
this is something cool. Do you have time to look at something?'
and you know I'm kind
of guilty of that a lot, I get in show-and-tell mode and like to share something recent
that's got me excited that I've come up with. - Yeah. You create something
that you really like
and you're like, 'Oh hey,
take a look at this.' - Mhm, mhm. So what kind of data does
your team collect and use? For instance, like what
is the volume of data that you work with and
how do you store the data?
- Well, typically I'm dealing
like order of magnitude on the hundreds of thousands of rows, as far as the data I'm ingesting. And it's usually in a similar amount as is actually being brought into Tableau.
The actual claims data,
it's mostly based on claims, that's based on is
actually billions of rows. So again, that data engineering team is pulling that together for us and making it more, just easier
to serve up analytics with.
So there's something kind
of funny that happens when you're programming, and I do ETL and stuff in SaaS usually, so when you get into those large data sets you can sometimes make
a mistake on a joint
and you end up having your job killed and you get an email that your job was killed with extreme prejudice. So that's kinda a right
of passage at Highmark, you have once you get that message.
- So once you and your team have collected and cleaned the data, you mentioned the data engineering team, what happens next and how
is that data being used? - Well, our primary customers are people
who have the title network
performance managers, and they are checking
how our members are doing and recovering in various
post-acute hospital facilities, like nursing homes or home healthcare. So we are looking at all those providers
and how the members
that are being cared for by those providers are performing, and the network performance
managers are looking at dashboards and helping
them see all that information and they are interacting directly
with those providers on
just performance, basically. - Cool. And what do the
data scientists produce from the results that you find like the dashboards, the written reports? I know you mentioned them.
Could you go in a little
more detail about it? - Let me dial it back
to what the customers are doing with the results. So the network performance managers, they may like compare one
provider to other providers
in the same region, or if there is say one, the
home health healthcare agency that have maybe 10 different
branches and one is struggling for some reason, they may dig in and do
some comparison of one
of the provider within
a large organization to the others and make a presentation with a subset of one of the dashboards and talk to the providers
about what's going on out there and try to get to the bottom,
you know try to take some
action and to correct that. - Ah, okay. So I know that you mentioned
a few specific software and programming languages that you use. Do you use them very often,
like Tableau, Excel, SaaS?
- Sure. Tableau, I am
in probably every day, almost every day. Whether
it's new development, maintenance, troubleshooting whatever. SaaS, I use kind of off and on. My SaaS program includes
a lot of prox equals
so heavy, heavy, SQL. And at the same time
I use Teradata Studio. So if I'm building a dashboard, what I like to do is create test scripts. So I have basically all the same results
in a SQL statement as close as possible and with a wear statement
that has all my filters in it. So I have that thing
going at the same time, I can look at a dashboard and compare it to results by hitting
the database directly
and see the difference. I also use other supporting tools, just like typical office stuff. And I'm a huge fan of Snagit as well for editing screen captures.
- Yeah, definitely. I know this question might be
kind of difficult to answer, but what technologies
would you predict to be most important to know for the future? - Yeah, that is a tough
question. Who really knows?
I think Tableau's in a very
strong position right now, and I see myself using it probably for the rest of my career. Like what does that mean,
10 or 11 years, you know? I'm older than a lot
of the users, but yeah,
it's gonna be around, and
the same with Power BI, I don't think they're going anywhere, Absolutely the same thing for SQL except it's been around longer, right? So I remember learning it in like
my junior or senior year in
college, which means 1989 or 90. So there, you know, it has staying power. To me, it's like math, you know. Math's not gonna go anywhere, SQL's not gonna go anywhere
as long as we had have structured data. I don't really use Python and R, but they are used a lot at Highmark. And I think either one of them or both is a good bet, and they
will continue to be used.
On the data vis side though, even if tools do change, the principles of data visualization aren't really gonna change
or based on human perception. So I think that's gonna keep going on.
Also, just in general business wide, I think one of the hardest
parts is gonna remain the same, which is human factors, like both my own and
those of my coworkers. So, and what I mean by
that is just stuff like,
you know 'how do you
tactfully deliver bad news' or, you know, 'are we
affected by cognitive biases in our decision making?', 'how well are we communicating
with the customers?' and you know, just conflict
and 'how do you prioritize?',
all those things. You know, they're gonna be around forever, you know, as long as people are around. - Yeah, completely agree. Especially on the math, but
I'm a little impartial to that.
- Oh, one more thing on that, the like the human factors thing. I have a article in my LinkedIn profile under my featured items called "Culture Eats Perks for Breakfast".
So it addresses some of those, I guess interpersonal work things and you know how people
handle those, you know in my mind being more important than perks like having a pinball
machine or something.
- Yeah. Yeah. Completely agree. So, could you talk about
a challenge that you or your team had to overcome
and how did you solve it? - That's a good question. Sometimes it's hard to
pick or quantify these
because like every day is a challenge. When you pick one, and
another one of my famous lines is if it was easy, it'd already be done. So, you know, we're typically trying to hit some hard problems.
But one that came to mind
from not too distant past was aligning goals across 6 dimensions, like time, lead type,
marketing campaign, county, and a couple other fields. So those goals versus actuals
and having the graph be
able to dynamically change based on how people filter. So that was in my prior position and how I solved it was I actually got some technical mentoring
from a very astute director that sent me in the right direction
and that really helped. More recently, I did a project involving a gap to goal calculation, which is measuring
basically a target rate
versus an actual rate and then the difference in the numerator of that calculation in
order to get the entity, whatever we're measuring, to that target. And so these were targets
across three different metrics
and they changed based on facility type. And I also set it up so that you could dynamically put your own goal in and it would recalculate
everything and color the outliers. And it was a pretty fun project.
I got excited about it and it was just cool to see it come off. And that's kind of one
of my mantras, you know, the value of this gap to goal calculation. I have a couple vises
out on Tableau Public
that demonstrate, and I
think it's broadly applicable for any entity that you
could be measuring that wants to hit some sort of target rate. - Yeah. That sounds
extremely interesting to me. I would like to maybe
use something like that
with my own students to
figure out their gaps as well. - Sure. I mean, check it
out. I'll send you the links. I have two different
dashboards that show it. One is in terms of NBA free throws. So like kind of rule thumb,
there is 80% is good, right?
So it goes out and kind
of compares players. I got some data from a 10 year period and shows like who was most efficient. And then I did in Tableau what's called LOD
calculation, level of detail,
to see which players basically
single handedly lost games because of their poor free throw shooting. So in other words, if they
would've hit 80%, right, the difference in the number of points they would've scored was greater
than the difference in
the loss of the game. - Oh. - Yeah. It's kinda, it's
a dubious distinction. - Yeah. That's cool. So what would you say is the
favorite part of your job?
- Well, I kind of get the
big three at Highmark: autonomy, mastery, and purpose. So autonomy, I have a lot of
say in the work I take on. I get to drive, somewhat, you know how we're making the dashboard,
there are standards but I
have a lot of say in that. And I can suggest new projects. And one of them came about, let me go off on a
tangent a little bit here. I saw one of our customers
presenting in a meeting
some of his findings from a dashboard. And what he had done is taken
two different screenshots and put them together on PowerPoint. And I said, well bells started going off in my head, and they
should for anyone, right?
If you see people taking screenshots and mashing together something, that should probably be a view that already exists in your dashboard, right? You need to make, unless it's
like a one off special case.
So I said, I asked him, you
know, "Is this important? Are you gonna do this
type of stuff a lot?" and they said, "Yeah, we are." I said, you know, "I
can make that for you." So I used Tableau Set
Operations to compare
basically two different markets along I think it's at least
12 different variables, you can slice and dice
them on simultaneously and check out all their metrics and display them down the
center at the same time,
kind of like when you see
a sports matchup, right, and they say, "what are
the keys to of the game?" and they show them in the center screen, they give a check mark on each
side who has the advantage. Well, I basically do the
same thing in a dashboard,
so that's a fun one. So that's all part of autonomy. Mastery, I like to think
I'm good at what I do. And I have chance to
take time to get better so I can attend, you know,
like one hour seminars
pretty much whenever I want
if nothing else is going on. You know, stuff like
going to a conference, like the virtual Tableau
conference, I, you know I need to arrange that
cause that's several days, but still, you know, that's encouraged.
And then the last pillar there is purpose and at Highmark, we envision a world where everyone embraces health, and I think that's a noble goal and one worth getting behind.
So, you know, those three
factors pulled together, yeah, make it a great job, you
know, I'm doing what I love. So it's really good. - Yeah. It sounds like it. I bet it's great to have that autonomy.
So if someone was interested
in a job like yours what education or training
would you recommend? - If you wanna get into
data visualization, build a public portfolio, right. Go out there and Tableau
Public, make some vises,
a picture's worth a thousand words, show people that you can do it, and that will help a lot. - Yeah, that makes a lot of sense. And could you recommend some resources
for our listeners to look into? - There's a lot of stuff out there. For Tableau specifically, you know, one of my big statements
is data vis and SQL can make a career. So just
those two together, right.
So check out my LinkedIn profile, I have an article there, it's the first one in
the featured content, "How I learned Tableau". For SQL, just one quick
one is w3schools.com.
You can go out there
to practice SQL there. - Yeah. I think our listeners
are greatly appreciative of those resources there. And were there any hard
skills or soft skills you needed in your career that
you didn't realize were important
when you were a student? - Well, again, when I was a
student was a bit of time ago. 35 years in computer science, I tried to reflect back on that. I'd say that were maybe two things
that I wish I would've known better. One, is a good understanding
of source code management, version control, and things like that. I knew nothing about that. And I think a lot of people
in analytics come out not knowing so much. So like everyone's using Git, so get out there and
learn that and understand how to keep your code in one safe place. Secondly, I'd say not really understanding
why software projects fail. I did have one software engineering class but I don't recall us
getting into the factors, the common factors, that cause
software projects to fail. And I would lump analytics in there
because there's a lot of overlap. So I encourage people
to just get out there, Google it, research it, you know. It's a little frustrating
to see these things happen over and over again
for like three decades,
and no one realizes that
these things are known and you can look out for
'em and try to avoid them. - Yeah. Maybe you could write
a little mini book on it. - Yeah. I at least have a post on it, but the top ones are kind of
comparing analytics and software projects. Soft skill wise, coming out of school, I was not very strong in that area. So I feel like I've come a long way. And again, I enumerated
some of those things in that
"Culture Eats Perks
for Breakfast" article. Some of the top ones
that I think, you know, were growth areas for me have been valuing and appreciating feedback. You know, and not taking
it personally and just,
you know, trying to make
whatever it is better to some sort of feedback process. Another one is like, if
you don't know something, admit it you know, right away, don't try to him and hall or guess at it.
Praising publicly and
criticizing privately. Double checking your own work before suggesting someone else is
the one making a mistake. If something's gonna be late, oh my gosh this is so
bad among developers,
if something's gonna be late, let people know as soon as possible. Like as soon as you
know it's gonna be late, that's when you need to fess
up that it's gonna be late. Something that just drives
people crazy is, you know
telling them on the last day or some people tell them after it's due. I wonder like, why didn't this thing you know, get in my inbox? 'Well, it's gonna be three more weeks',
you know you don't wanna
be in that position. So things like that have
been growth areas for me. I'm in a lot better position
than I was 30 years ago. You would hope with three decades, one would make some progress.
So there you go. - Yeah. Those are
excellent, excellent points. And I hope our listeners are
definitely listening to that. You're quite on active on LinkedIn. And I know you talked
about it quite a bit,
but could you tell us a little bit more about how you utilize the platform? - Sure. To me, it's all about helping young analysts with tips. I mean, and that's it.
I do a lot of posts about Tableau. I think I have a unique voice out there because I combine that with
other factors about analytics, about SQL, about getting quick with everyday tools that you
use, like the Office Suite,
and really that's what
my content is about. And I started during the pandemic and just kind of stuck with it. I focus on introductory
advice and kind of just like if something goes wrong
for me during the day,
or something's not obvious and I think it would be helpful
for other people to know it, that's what I write about. - Finally, before we let you go is there anything else that you'd like
our listeners to know that
we didn't cover today? - Oh, there is a lot of
stuff I could talk about. So it depends on like what
rabbit hole you want to go down. One is, this is another thing that's been known for 25 years,
that limiting the interruptions
of knowledge workers can make a huge difference
in your productivity. Not just a little difference, I'm talking about doubling
your productivity. So if you can, at work,
carve out two 90 minute sessions where you are completely
uninterrupted: phone down, notifications off, you
are not looking at email, you're not being pinged by Teams, none of that, right?
You can get a ton more work done. And that's because knowledge workers take 20 minutes to reach top speed, right, to get in a full state of flow. So limit those interruptions
and it will make a huge
difference in what you get done. And I have a graph in my
Tableau public profile that shows the difference in interruption on two simulated days, on one with like random interruptions
and one with these two 90 minute blocks, and how much more you can get done if you allocate those for yourself. - Great, thank you so
much for joining us today. And to all of our listeners,
if you haven't already checked
out our previous podcasts, they are available wherever
you listen to podcasts. For more information about starting a career
as a data scientist go to www.dataanalytics.buffalostate.edu.
Don't forget to subscribe so that you get notifications each time we release a new episode
of Buffalo State Data Talk. (exit music)
Some content on this page is saved in PDF format. To view these files, download Adobe Acrobat Reader free. If you are having trouble reading a document, request an accessible copy of the PDF or Word Document.
