September 4, 2024 • 44 mins
Dr. Julie Alig, known as 'The Data Diva,' shares her insights into the transformative potential of AI, drawing on her extensive experience with quantitative tools and machine learning. They discuss practical applications, the current limitations of AI, and its potential for future impact, particularly in data analysis. The conversation also touches on the anxieties and barriers people face when dealing with data, emphasizing the importance of visualization and understanding data's broader implications. The episode highlights the importance of a cautious yet optimistic approach toward integrating AI into meaningful and ethical work.
To delve deeper into data and AI, visit:
- Dr. Julie Alig’s Website
- Connect with Dr. Julie Alig on LinkedIn
- Join the data discussions on Friday afternoons (3 PM Eastern Time) on LinkedIn or catch the recordings on her YouTube channel.
For insights into utilizing AI for positive change, follow Sharon Tewksbury Bloom’s work at Bloom Facilitation. If you have suggestions for future podcast interviews, contact Sharon. Stay curious, stay informed, and explore the myriad ways AI and data can transform our work and lives.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Brian AI (00:01):
Are you not sure how to feel about the way AI is suddenly everywhere?
AI for Helpers and Changemakers isa show for people who want to do
good work and help other people.
Whether you're already using AItools and loving it, or you are
pretty sure that ChatGPT is thefirst sign of our downfall, we want
you to listen in and learn with us.
(00:22):
Your host on this journeyis Sharon Tewksbury Bloom.
For 20 years, she's workedwith helpers and changemakers.
She believes that we're about to seethe biggest changes in our work lives
since the Internet went mainstream.
We're in this together.
Join us as Sharon interviews peoplein different helping professions,
navigate what these new technologiesare doing to and for their work.
Sharon (00:44):
Welcome, Julie.
Thanks so much for coming on today.
I would love it if you would justintroduce yourself to our listeners
Julie (00:50):
It's my pleasure to be here.
Thank you for the invitation.
my name is Dr.
Julie Alig.
I go by the moniker, The Data Diva.
And, how have I come about, learningabout AI and using it so much?
I learned way back in graduate school whenI was getting my doctorate, I learned a
(01:12):
lot of, quantitative tools and machinelearning and a lot of those are subsumed
today into what is referred to as AI.
And so I've been doingthis for a while, but.
(01:33):
The difference is thegenerative part of it.
So the generative AI, that's verynew, and I'm right there with
everyone else learning how to useClaude ChatGPT Gemini and others.
Sharon (01:49):
generative AI is really the new kid on the block.
It's the exciting thing that's bringingin a lot of new possibilities, putting all
of us trying to learn as much as we can.
How is generative AI changingwhat's possible right now
with data and with your work?
Julie (02:09):
It's the sort of thing that
it's such a change and such atransformation that I don't even know
how much of a change it's going to end uphaving been, if that kind of makes sense.
I heard, I can't remember who saidthis, but it's like the internet, sure.
(02:33):
It's like electricity.
When electricity was discovered,people had no idea, they couldn't even
conceive of what would be possible.
It was that profound a transformation.
I think this is goingto be on par with that.
(02:57):
I also think we are very far.
away from that being our reality.
right now for where generative AI is,there's still a lot of hallucinations.
And you still cannot expect
100 percent veracity, or even sometimes50 percent veracity from things, if
(03:23):
you're asking it to create subjectmatter or something like that.
What it is really good at is, andwhat I'm finding myself personally
and my clients are finding, isbeing able to automate those rotes.
tasks that we do over and over thatespecially with, no code or low code,
(03:50):
things like, like Zapier or make orsomething, those are really going to help.
that's one of the big ones.
I think there aren't enoughpeople who really are aware.
of all the possibilitiesgoing on with that.
And again, going back, I think there'sa part of this that, that is the
(04:11):
new toy, the shiny new object thing.
But I'm also hearing a lot of peoplebeing intimidated by this because
again, it's math and it's robots, right?
a lot of these no code, low codesolutions, I think once they become a
little bit more broadly used, not just atthe enterprise level or the Gen Zers or,
(04:36):
and Millennials, I think people are reallygoing to be surprised at the efficiency
and productivity gains that they get.
as far as me, there are are people outthere, some of the more well known people
at the cutting edge of research on AI,the generative AI, who are coming up
(05:01):
with these examples of having in, sentthe generative AI a data set and told
it to analyze it and it's come back.
And I, it'll come back, sure, I,I don't trust it, farther than I
can throw it and not even there.
(05:22):
And I think people are,
that is rather dangerous right now.
So when I use it in my dayto day work, I'll use it.
I use, I code.
To do my analyses in Python and inR, and if I can't remember the code
for something, I'm able to say, I'mtrying to do this, give me the code
(05:45):
in R, and it usually gets it right.
It doesn't always.
It doesn't always get just plainold coding correct, much less
interpreting what the output says.
I don't use it for that at all.
Even though there are other peoplewho are out there saying, Oh, it
can analyze your data for you.
(06:05):
It can not.
It'll be great once it can, but for rightnow, you still need to have that knowledge
and understand what a p value is.
Understand when you tell it to handlethe missing values, in your data set.
(06:25):
And it says, Oh, great.
I just got rid of them all.
Or I did.
pairwise deletion.
Why the heck did you do pairwise?
There's nothing paired in here.
We're not, that sort of thing.
So some of the more detailed sortof things, you really need to
know and you need to be checking.
But, so as far as in my day today work, helping me to also,
(06:52):
craft proposals End for my clientsfor the kind of work that I do.
I have done enough of them now after fiveyears that I know what I want to say.
I know what the important pointsare going to be and I am very
good at using the prompts.
(07:14):
to be able to come up with somethingmuch more quickly than I could
if I were just working by myself.
I think one thing to keep in mind is,
and I hadn't even realized this untila friend of mine mentioned it, was,
a lot of people think that thesegenerative AI models are like Google,
where you want to put in the fewestamount of words and they're 180 degrees.
(07:41):
You want to be as rich and detailfilled, providing as much context
as you possibly can in order toget back something really useful.
really getting used to being quite wordyin your prompts and quite directive,
(08:01):
giving it examples, so zero shotor one shot or many shot prompting.
those are going to be the key, for othersas we all get used to this new tool.
But those are the things that arereally helping me the most right now.
Sharon (08:20):
Yeah, I think that you're echoing something I've heard from
other guests, which is that right nowit feels like it's really great as an
assistant for you when you know how todo something and you can give it clear
directions and you can check its work.
that's something that it can do well interms of being able to speed up your own
(08:42):
process or help you with some parts ofit that are already pretty well dialed.
Or even as a thought partner, one thingthat I haven't tried, but now I'm curious
to try would be instead of giving it thedata set and asking for it to analyze
it, I'm curious if I could give it thedata set and say, what are the types of
(09:02):
analysis I could do on this data set?
like what kinds of, questions couldI answer or, so it'd be fun to cause
I do the same sort of thing in mywriting, where sometimes I'll say,
okay, this, these are some ideas I have.
Could you offer me five differentmetaphors that I might be able
(09:24):
to use to take this idea further,and use it as a thought partner.
So it'd be interesting to get it to just.
Remind you of Oh yeah, have youthought about presenting it?
It's this kind of graph, or haveyou thought about, looking at this?
because I have actually, I have put adata set into chat GPT and said, can
you tell me, some basic things aboutthe data that I knew I could, check
(09:50):
the answers because I was like, will itbe able to pull this information out?
It was accurate, but I was asking itvery simple things in terms of like,
how many people on this list are fromArkansas, or convert this to a table
where it shows what state, how manypeople per state, things like that.
So it could do some very basic things.
(10:10):
and that is something that I hope willbe possible in the future, is that
for people who just want somethingbasic, take my spreadsheet and turn
it into a table that I can put in myreport that shows the same information
and a slightly different visual.
It will make the barrier to entry for thatsmall amount of use of data better, so
(10:31):
more people feel more comfortable with it.
But right now we do have to be ableto check the work on everything.
Julie (10:39):
Everything.
Yeah, I hope we're there, too.
I really do.
And yeah, the same thing, I tellpeople, treat it like a summer intern.
They're really raring to go.
They want to please you.
They do not want to admitthey're making mistakes.
And you have to check everything.
But they're very nice andthey're very, yes, exactly.
(11:03):
But, but just with that, though.
Just with what you andI have talked about,
the possibilities, the next year,I just, I can't even imagine
what's coming down the pike for us.
So that's why I remain optimistic,cautiously optimistic, because we
(11:24):
haven't even touched on all theethics and the bias around all
of this.
but cautiously optimistic, withwhere this is headed and how it
can help us in our day to day.
Sharon (11:37):
Yeah.
I think the huge amount of variety ofways that could be used in the future.
One thing that's exciting, and Iguess I'll save this and then pivot
it to maybe a final question for you,which is I have been really enjoying
learning about ways that people arecombining technologies right now.
They're calling it multi modal.
(11:58):
So for instance, a, robot math tutorwho's able to have the student hold up
what they wrote on the paper and theycan actually, Take in that image and
know that the student wrote two plustwo equals five and then know that's
not correct and be able to, prompt back.
(12:21):
So being able to combine multipletechnologies, Is something that
just infinitely increases whatthese tools are capable of.
I'm curious What's something that you'refollowing closely or really loving
learning about right now that's on theedge of, what you've been hearing about
that you want to learn more about?
Julie (12:42):
Definitely the multimodal is on my radar.
I was at a, at conference lastmonth down at MIT about AI.
and it was all over.
It was all over the place.
So it's an area I have not been paying awhole lot of attention to prior to that.
now I'm thinking more.
(13:05):
And it's just, you think ofthese as a language model.
And so you think language, written.
But you're not thinking aboutvisual, oral, all those others.
And it's just.
All of a sudden, it's like you're walkinginto a completely new room in your house.
(13:26):
it's I had no idea this was here.
This is cool.
I always wanted a room like this.
so multimodal is definitely one of thethings that I'm keeping my eyes on.
And then what I was talking about earlier,the low code, no code solutions, there's
a lot more of those that are coming.
(13:46):
coming across my radar, I think thatif we're going to be using these tools
in anything beyond something like writemy 500 word essay or, something like
that, I think those are going to beone of the ways for people to really
(14:09):
be able to reap some serious benefits.
And then I'm always keeping my eye onwhat's going on with the data analysis.
It's just not there yet.
It's just not.
but I think it will be.
So I'm very eager to get some,some good evidence of that.
Sharon (14:29):
I heard someone lamenting the fact that they had found that
the tools were only as intelligentas a human three year old.
And I was like, wait,I've met three year olds.
Some of them are very intelligent.
that seems like we should bethinking, Wow, they're already as
intelligent as a three year oldand we're just getting started.
Julie (14:50):
No, I'm with you on that.
Sharon (14:52):
This is your host Sharon Tewkesbury bloom.
And I just want to pause brieflyto say that we really appreciate
that you've been listening to AIfor helpers and change makers.
I am supporting organizations who aretrying to figure out how to utilize AI.
Through my work at bloom facilitation.
So, if you are interested in howto bring new practices into your
(15:17):
workplace, I hope you'll reach outto me through bloom facilitation.com.
Now for more.
With my conversation with Julie.
many listeners, feel like AI is brand new.
They're racing to catch up.
They feel like they justheard about these tools.
but I'm a history majorby original discipline.
And so I love Figuringout, how did this evolve?
(15:39):
Where did this come from?
What were the steps that came before this?
And so I've also learned a bitabout the different waves of
technology advancement in this area.
And like you said, machinelearning is one of those.
so can you give us maybe a little bit ofbackground about what you had studied.
Julie (15:59):
Sure, it's interesting you mentioned you were a history major.
I was a French literature and historydouble major, and look at where we are
now, doing all of this techie stuff.
my head is in numbers all day,every day, and I love it, machine
learning has been around for a while.
it's a way of telling machines what to do.
(16:24):
similar to something likecomputer programming.
If you think back to the early days ofcomputer programming, you would, have
punch cards that would get fed in andwhatever you told it to do was output.
Machine learning was a step further it wasmore descriptive, more prescriptive to the
(16:46):
machine about what you wanted it to do.
And so the models were.
What I got in when I was in my doctoralstudies was with the quantitative
analysis part of it, which is heavilyinfluenced by statistics and econometrics.
So yeah, not everyone's goingto have any idea what that is.
(17:09):
It's basically using computers,using machines to help find patterns.
in a very high level way what machinelearning is about a part, under the
umbrella of artificial intelligence,that is one very well defined field
(17:35):
or subdiscipline that feeds rightinto everything else that's going
on today with those generative AI.
Because what is generative AI?
It's actually generating.
So traditional machine learning, typicalmachine learning is, it's not generating
(17:56):
anything based on what I tell it to do.
the generative part of it.
is like we see with chat GPT,with Claude, with some of these,
the language models, right?
It's actually generating paragraphs,output, conversations, that sort of thing.
So you can get into that muchmore fluid back and forth.
(18:20):
That's not typical machine learning.
how does it generate all this?
It goes back to prediction.
these large language models hooveredup all the text from the internet,
running equations and algorithmsto figure out what words were
(18:42):
most likely to beassociated with other words.
And so once you know that, then youcan start predicting what the next
word is going to be and So it's usingthat high level kind of thinking of
like machine learning But it's puttingit to use with predicting words and
(19:04):
the there's more to it But that'sbasically how they're tied together.
They build on top of the other.
Sharon (19:14):
that was really eyeopening to me.
I just completed this MITexecutive education course around
AI, and I didn't realize thatso much of this is statistics.
So much of it is based on probabilitiesand on being able to notice and
learn from patterns, then recognizepatterns, anticipate patterns.
(19:38):
I was not very good atstatistics, I'll admit.
I was really good at algebra.
I was really good at other kinds of maths.
Statistics was the bane ofmy existence in high school.
to know that these tools I'm excitedabout are based in statistics is amazing.
Julie (19:54):
Don't give up hope.
My first time with statistics in gradschool was nothing to write home about.
it wasn't, absolutely.
It's a different way of thinking.
Sure, it's math.
I'm good at algebra.
This is just a different way of thinking,just like calculus is a different way,
(20:15):
and you got to wrap your brain aroundthat, and it doesn't always happen
neatly within the confines of onesemester but I would encourage you, or
anyone else out there who's curious,keep at it, and you'll be surprised
by how you're able to grasp things.
Thank you.
Much more than you thought.
(20:36):
That's what I tell my students, too.
And it works.
It works.
They do get it.
But, yeah, it's all probability.
These generative AI, large languagemodels, it's all probability.
And that's wonderful, but that'salso something that we need
to keep in mind is a downside.
And it can be dangerous.
(20:56):
Because the same sorts of things that,The same caveats that I have to talk
about when I do forecasting and thosesorts of things for my clients and,
in my own work, I only have the pastdata, the history, to base it on.
And the same thing withthese large language models.
(21:19):
They're not necessarily built.
or good yet at conceptualizingsomething that hasn't happened.
And so they also can fall into thetrap of almost seeming like they
have blinders on because it's simplythe data that was fed into them.
(21:41):
They're not able to think outsideof that, at least not yet.
And that's where, humans, we reallyhave a leg up on the machines,
Sharon (21:53):
And I think for something relatable for a lot of people in terms
of how these tools and algorithms wereoriginally brought into many people's
lives is something like Netflix andhow they recommend movies to you
and how, based on your own viewinghistory and maybe what you spend a lot
of time looking at in the previews,they start to recommend other movies.
(22:17):
And, I do feel like I fight withthe machine sometimes where I'm
like, Don't put me in a box.
that's not the only kind ofthing I contain multitudes.
I like sports documentaries and Ilike cheesy romances and I like,
Nazi Germany history movies, becausethat was my, studies growing up.
So I'm sure Netflix is very confusedby me, but it's also a great example
(22:40):
of how, the machines and computersare trying to make sense based
on the data they already have.
And they're trying to notice patternsand predict things, but they can only
work with the data they already have.
They don't know that I've gonethrough some metamorphosis and I'm
suddenly interested in somethingbrand new that I've never shown
(23:01):
any inkling towards before.
Julie (23:05):
exactly.
They're advanced pattern recognition,trend recognition tools, if
we think about it that way.
Think about on your YouTube Timelineand every so often you'll get something
that pops up and says, do you wantsomething completely different?
And that's exactly so YouTube, atleast this is what I'm assuming.
(23:31):
They're trying to be ableto pull us out of that.
What might be a trap almostnot a trap, a dead end, maybe.
Maybe.
Sharon (23:39):
Yeah, I joke that my dad has reached the end of YouTube on his
subspecialties because he has somevery narrow interests that I think
he has watched every single thing
Julie (23:50):
Yeah.
Sharon (23:50):
in that.
If it's about earth moving machinesand it's been recorded by a bearded
white man, he has watched it.
So, and he knows because I make funof him about that, but All right.
So you are the data Diva.
At some point, I'd love to hear ifyou gave yourself that title or if
someone else gave you that title.
(24:11):
And then I'd also love to hear howdata, what you think people who
don't study data all the time needto know about data to, live in this
world today with how it's beingshowing up in everything we're doing.
Julie (24:28):
Mhmm.
I'll take that first question first.
my husband came up with it andhe came up with it years ago.
back when, I was in oneof my full time jobs.
And, when I went out on my own,doing my own consulting, my analytics
consulting, at some point I justjokingly said that in a networking
(24:51):
group and people really liked it.
And I was like, Oh, Iguess my husband was right.
so it comes from my husband, my
better half, I call him, and whatdo people need to know about data?
um, one research study recentlyfound that 93 percent of Americans
had some sort of math anxiety.
(25:15):
And at least in the businesscircles, the business ecosystems
where I am right now, a lot of timesdata equates to dollars and cents.
Equates to something that my accountantdoes for me, or my bookkeeper, or my CFO.
(25:39):
They handle that so that Idon't need to look at it.
And, my question back usually issomething like, You realize that
mountain of data has gold in it.
And your accountant or your CFO, theymight be looking in certain places, right?
(26:04):
For certain things so thatthey are able to create your
profit and loss sheet, right?
Your quarterly reviews.
There's a whole lot more in that mountainof data that if you had a different tool,
like some statistics, you be able to minethat for gold and diamonds and emeralds.
(26:29):
And I'll usually say something like that.
And then I'll give a couple of examplesabout because a lot of people think
my business doesn't have any data.
And I have clients that all they hadwas their financials, I was able to
find underperforming profit, profitable,offerings, I was able to help them to
(26:54):
untangle, what was going on with customerretention and why having their sales
team just continually bringing in morenew customers wasn't moving any needles.
And so with the data and the kindsof tools that I use, these more
statistical tools, I'm able to saywhy these things are happening.
(27:18):
And once you know why they're happening,then you can start fixing them.
And ultimately, businesseswant to stay in business.
They want to be profitable.
They want to be sustainable.
And many businesses, business owners,want to be able to have a legacy.
(27:42):
And If you're just drowning in redink and you don't know what's actually
making money, what's actually drivingyour customers to your competitors
or to just leave, which marketingchannel is actually working, you're
never going to be able to achieve that.
Wow.
Sharon (28:07):
fighting the good fight on that.
it made me think of this businessthat I know about, where I actually
had one of my first jobs ever.
It's a locally owned, familyowned retail business.
They just celebrated 50 years in business.
The daughter was able to takeover ownership from her father.
(28:30):
So she's now looking to thefuture of the business, and she
cares a lot about sustainability.
And so not only is she interested incontinuing to run a profitable business,
but because they do have this stablefoundation, she's now thinking, how
could we do even more for our community?
(28:51):
How could we do more for theenvironment or have a lesser
impact on the environment?
And I'm curious beyond just the numbersand cents, what other ways might
there be gold in those data hills forsomeone like that who's looking at
that sort of, what do they call it?
Triple bottom line business wantto do good for customers, the
(29:17):
environment and their business.
Julie (29:19):
I'll go back to what I say a lot of times to the skeptics
who say, my bank balance is fine.
My employees are paid.
I don't have any fines from the IRS.
What do I need to worry about?
And yeah, you don't needto worry about anything.
(29:39):
if that's your goal, then you're right.
You're good.
If you want to actually do somethingand you want to, be a good citizen in
your community, you want to be a goodsteward of the environment, anything
like that, you're not going to be ableto if you don't have a solid foundation
(30:07):
of what your business is doing wrong,what you can improve on in those.
it's all well and good tosponsor, Arbor Day or something.
Is that really making a difference?
And if your business isn't makinga profit, are you going to be
(30:30):
able to, sponsor any of that?
If your employees aren't engaged,healthy, both mentally and
physically, fulfilled with their jobs.
Are they going to be willing to, evenif you pay them for the day, right?
Sharon (30:48):
Yeah, and maybe part of what I was thinking is.
If you don't know, for instance, yourenergy usage as a business and how that's
also related to your costs, like how muchare you paying for heating and cooling
and, that's often a great place to startwith something like improving your impact
(31:08):
on the environment because there couldbe that win of this could lower costs
as well as be able to use less energy.
But those are data.
those are sources of data of being ableto track what are your energy uses?
How does that relate to yourability to run your business?
(31:28):
Are you staying open later hoursbecause you want to run a promotion,
but then does that cost youmore in energy or, other things?
So it's just interesting allthe layers that I can see of
potential data inputs to that.
Julie (31:43):
Yeah.
And if you've also got a solidaim in mind, it makes it easier to
convince people to track your data.
So no one likes tracking.
no one likes tracking anything.
I don't like tracking, did Ibrush my teeth this morning?
(32:04):
Did I brush it after lunch?
Anything like that.
And I see so many in our space outthere that almost like admonishing.
business owners andbusiness professionals.
Like you need to be tracking this.
You need to be tracking that.
Coming up with these checklists, theselaundry lists of things to, to track.
(32:27):
And I'm like, yeah, sure.
We can track everything from here tothe moon, but does it make a difference?
And is it going to really, move the needleon the things that our business needs to?
If instead, like what you said, Sharon,We have a commitment to reducing our
energy usage, then boy, someone isgoing to be on top of that, figuring
(32:53):
out what it is on a day by day basis.
today when we're recording this,I'm in the Northeast and it's
very hot, unseasonably hot.
Energy usage is way high.
What if I'm a business owner?
If I'm a manufacturer, whatam I going to do to offset?
all this extra energy that Ineeded to expend or buy today.
(33:17):
I have no idea, again,if I'm not tracking.
So it's a nice way also to really makeit clear to the members of your team.
Like this isn't justputting numbers in boxes.
We're going to do something withthis and you've got a role to play.
Yeah.
Sharon (33:40):
three possible barriers or hesitations people might have that
there's math anxiety that keeps peoplefrom digging into their numbers.
Perhaps they might be skeptical aboutthe benefits they'll get from actually
collecting tracking and using the dataand then tracking itself can be a chore.
(34:02):
is there anything else that you've seen?
that kind of keeps people fromdiving into their data and getting
the most benefit out of it.
Julie (34:13):
People don't draw enough pictures.
and by that um, what's the linguafranca of business in the world,
not just the United States, whenwe think about numbers and data.
It's Microsoft Excel, GoogleSheets, something like that, right?
(34:34):
And what is that?
It's columns and columns of numbersand rows and rows of numbers.
And if you don't put the format inthen it's all those ugly, number
signs in there that some peoplehave no idea how to get rid of.
And it's just, it's overwhelming.
even me, I don't know about you, but evento me, some of these are like, I just
(34:57):
feel like I'm being drowned in numbers.
And I can't blame anyoneelse who's, doesn't have fun,
like me, looking at numbers.
I don't blame them for not being all gungho about digging in and what have you.
The one thing that, I use with myclients, I use with myself, I use with my
(35:22):
students is draw a picture, make a graph.
doing that, the human brain, just theway that the human brain is hardwired,
it understands and perceives visualstimuli much more quickly and in a lot
(35:42):
of ways much more completely than othertypes, than, looking at numbers on a
spreadsheet, or even hearing things.
and so sometimes, most of thetime, the first thing I do is
start making, drawing pictures.
I make bar charts, scatterplots, and line graphs.
(36:07):
Those are the three that researchhas shown the human brain
is, best at perceiving with avery high degree of accuracy.
And so bar charts are not sexy.
They're not necessarily fun, but if you'vegot like several hundred thousand rows
of data and we've all been there, you canuse a bar chart and it will start cutting
(36:31):
through all the noise and it'll startshowing you, Oh, that's what's going on.
And then you can go from there ormaking a line chart or something.
honestly, I really do think making anExcel makes it easy enough, to simply
insert and it's got, a couple ofpoint and clicks and it'll make them.
(36:54):
It's a completely separate topicif they're actually necessarily
any good, the defaults that arein there, and that sort of thing.
But at least it's a way to cutinto the data and start to see
what the heck is going on here.
Sharon (37:12):
Yeah.
I have two follow up points on that.
One is also, sometimes it's not until youtry to make that picture That it requires
you to really ask yourself, what questionam I trying to answer with this data?
(37:33):
Cause you are asked like, whatdo you want to see on those bars?
What is your X axis?
What is your Y axis?
What do you want to see compared to what?
and I know I recently was working withAmeriCorps programs on recruitment.
And one thing we often ask them is,where are you getting people from?
Where do people find out about you?
(37:54):
That's important to know, but a lotof people have stopped at that level.
So they'll just have maybe a piechart that shows, okay, we got
this many through Indeed, this manyfrom a referral from a member, this
many from a job fair, et cetera.
But then I've asked them, okay,that's one piece of data, but is
(38:15):
that really what you want to know?
Or can we go deeper and say, Whichof those referral sources actually
led to members who didn't just applybut actually became members and even
better yet, if we have the data, whichof those actually were successful as
(38:36):
members and made it through to the end.
oftentimes, No one's ever gonethrough that data with them and
said, we could go deeper on this.
We could find out some reallyinteresting answers to our questions.
Not just the first default, which iswhere did the applications come from?
Julie (38:54):
yeah, exactly what you were talking about.
Lead sources and conversion rates.
That kind of pipeline analysisis something that I do for some
of my clients, and it's eyeopening, absolutely eye opening.
When we look at, the sources thatmight have the largest volume.
(39:18):
are not necessarily the oneswhere we get the stellar whatever
kind of customer or something.
Yep.
And it's very easy to show thatthen with a, some sort of visual.
Yep.
Sharon (39:35):
Excellent.
Yeah.
I also had done a training back in theday called making friends with metrics.
Okay.
Because, as you've found also, peopledo have a lot of anxiety about numbers,
and I work with a lot of groups throughuniversities, through grant funding,
etc., and they're required to trackdata, and they're required to report
(39:56):
it to a funder or to a higher up, butthey're not using it for themselves.
They're afraid of it.
And trying to help them
be empowered to actually say,this is your data, you can use it.
to help you feel more confident indecision making or to understand
where to put your resources on timeor to advocate for more funding
(40:19):
or whatever it is that you need.
and one example I show because it's ofteneasy for people to picture is that I used
to run the litter collection program,adopt a street, adopt an avenue, and so
we would do a very rudimentary type of.
data collection, which is that we'dhave a little slip of paper on the
(40:40):
bucket that the volunteers tookout with them to pick up trash.
And they would just record how manyvolunteers for how many hours picked
up how much, how many bags of trash.
So it wasn't fancy.
We weren't using any tech.
It was just a pencil and a littlepiece of paper, but then I could
input that into my spreadsheet.
(41:03):
And then we did actuallyhave a fancy GIS map.
And we were able to start overlaying thetrash, volume data over the map of the
city and we could start to see where'sthe higher concentration of litter.
In the city, geographically.
(41:26):
And then that led us to be ableto make better decisions about if
we have a big cleanup day, likeon earth day, are we going to send
people to the most popular park?
That's also the one that never needsany trash picked up because everybody
loves that park and they pick up theirtrash and keep it clean to begin with.
Or are we going to send them, to maybea part of the neighborhood that has.
(41:49):
It's, a lot of busy commercialarea with dumpsters that get flown
open and, everything gets broughtby the wind down to one low point.
And so all of the trashgets collected in that area.
so you start to see those patternsthat show up and it's just so
visual because it's literallylitter and it's in your city.
(42:10):
it helps people understand, Oh yeah,that would be really nice to be able to
know, like, where's the litter and thenwe can start to investigate why is it
always showing up in the same places andthat metaphor is the same for business
or same for people's operations of like,Where are the things showing up all the
time and why are they always showing up?
Julie (42:31):
Exactly.
Sharon (42:32):
thank you so much for Lending some of your insights
and wisdom and knowledge with us.
How can people learn more aboutyour work and maybe even contact
you if they have their own data thatneeds a diva to take a look at it.
Julie (42:49):
You got it.
I have a website, www dot janalytics, all one word.com.
If you go there, youcan book a call with me.
You can download, I've got a couple of,Free resources there to help professionals
(43:11):
and business owners try to untangleand make some sense of their data.
I hang out over on LinkedIn, so definitelycome over there, find me, connect with me.
Would love to meet you over there.
and I host live streams.
On Friday afternoons where Italk about something to do with
(43:35):
using data and AI in business.
So it's usually three on Eastern time.
So noon Pacific, early Saturdaymorning, in Australia, New Zealand.
I've got a few people overthere who watch, so I would
love to see anyone over there.
(43:57):
And then, pretty much.
Sharon (44:00):
Awesome.
Yes, I've been able to tune intosome of those lives and they're very
insightful, so I'd highly recommend them.
You can watch them live, or Ibelieve they also show up if you
can see the recording afterwards.
Do you know if that's true?
Julie (44:15):
You can, you can see the recording over on YouTube channel
at nhdatadiva.
There you go.
Sharon (44:23):
Awesome.
Brian AI (44:24):
Thank you for joining us on this episode of AI for
Helpers and Changemakers.
For the show notes and moreinformation about working with
Sharon, visit bloomfacilitation.
com.
If you have a suggestion for whowe should interview, email us
at hello at bloomfacilitation.
com.
And finally, please share thisepisode with someone you think
would find it interesting.
(44:45):
Word of mouth is our best marketing.
Are you not sure how to feel about the way AI is suddenly everywhere?
AI for Helpers and Changemakers isa show for people who want to do
good work and help other people.
Whether you're already using AItools and loving it, or you are
pretty sure that ChatGPT is thefirst sign of our downfall, we want
you to listen in and learn with us.
(00:22):
Your host on this journeyis Sharon Tewksbury Bloom.
For 20 years, she's workedwith helpers and changemakers.
She believes that we're about to seethe biggest changes in our work lives
since the Internet went mainstream.
We're in this together.
Join us as Sharon interviews peoplein different helping professions,
navigate what these new technologiesare doing to and for their work.
Sharon (00:44):
Welcome, Julie.
Thanks so much for coming on today.
I would love it if you would justintroduce yourself to our listeners
Julie (00:50):
It's my pleasure to be here.
Thank you for the invitation.
my name is Dr.
Julie Alig.
I go by the moniker, The Data Diva.
And, how have I come about, learningabout AI and using it so much?
I learned way back in graduate school whenI was getting my doctorate, I learned a
(01:12):
lot of, quantitative tools and machinelearning and a lot of those are subsumed
today into what is referred to as AI.
And so I've been doingthis for a while, but.
(01:33):
The difference is thegenerative part of it.
So the generative AI, that's verynew, and I'm right there with
everyone else learning how to useClaude ChatGPT Gemini and others.
Sharon (01:49):
generative AI is really the new kid on the block.
It's the exciting thing that's bringingin a lot of new possibilities, putting all
of us trying to learn as much as we can.
How is generative AI changingwhat's possible right now
with data and with your work?
Julie (02:09):
It's the sort of thing that
it's such a change and such atransformation that I don't even know
how much of a change it's going to end uphaving been, if that kind of makes sense.
I heard, I can't remember who saidthis, but it's like the internet, sure.
(02:33):
It's like electricity.
When electricity was discovered,people had no idea, they couldn't even
conceive of what would be possible.
It was that profound a transformation.
I think this is goingto be on par with that.
(02:57):
I also think we are very far.
away from that being our reality.
right now for where generative AI is,there's still a lot of hallucinations.
And you still cannot expect
100 percent veracity, or even sometimes50 percent veracity from things, if
(03:23):
you're asking it to create subjectmatter or something like that.
What it is really good at is, andwhat I'm finding myself personally
and my clients are finding, isbeing able to automate those rotes.
tasks that we do over and over thatespecially with, no code or low code,
(03:50):
things like, like Zapier or make orsomething, those are really going to help.
that's one of the big ones.
I think there aren't enoughpeople who really are aware.
of all the possibilitiesgoing on with that.
And again, going back, I think there'sa part of this that, that is the
(04:11):
new toy, the shiny new object thing.
But I'm also hearing a lot of peoplebeing intimidated by this because
again, it's math and it's robots, right?
a lot of these no code, low codesolutions, I think once they become a
little bit more broadly used, not just atthe enterprise level or the Gen Zers or,
(04:36):
and Millennials, I think people are reallygoing to be surprised at the efficiency
and productivity gains that they get.
as far as me, there are are people outthere, some of the more well known people
at the cutting edge of research on AI,the generative AI, who are coming up
(05:01):
with these examples of having in, sentthe generative AI a data set and told
it to analyze it and it's come back.
And I, it'll come back, sure, I,I don't trust it, farther than I
can throw it and not even there.
(05:22):
And I think people are,
that is rather dangerous right now.
So when I use it in my dayto day work, I'll use it.
I use, I code.
To do my analyses in Python and inR, and if I can't remember the code
for something, I'm able to say, I'mtrying to do this, give me the code
(05:45):
in R, and it usually gets it right.
It doesn't always.
It doesn't always get just plainold coding correct, much less
interpreting what the output says.
I don't use it for that at all.
Even though there are other peoplewho are out there saying, Oh, it
can analyze your data for you.
(06:05):
It can not.
It'll be great once it can, but for rightnow, you still need to have that knowledge
and understand what a p value is.
Understand when you tell it to handlethe missing values, in your data set.
(06:25):
And it says, Oh, great.
I just got rid of them all.
Or I did.
pairwise deletion.
Why the heck did you do pairwise?
There's nothing paired in here.
We're not, that sort of thing.
So some of the more detailed sortof things, you really need to
know and you need to be checking.
But, so as far as in my day today work, helping me to also,
(06:52):
craft proposals End for my clientsfor the kind of work that I do.
I have done enough of them now after fiveyears that I know what I want to say.
I know what the important pointsare going to be and I am very
good at using the prompts.
(07:14):
to be able to come up with somethingmuch more quickly than I could
if I were just working by myself.
I think one thing to keep in mind is,
and I hadn't even realized this untila friend of mine mentioned it, was,
a lot of people think that thesegenerative AI models are like Google,
where you want to put in the fewestamount of words and they're 180 degrees.
(07:41):
You want to be as rich and detailfilled, providing as much context
as you possibly can in order toget back something really useful.
really getting used to being quite wordyin your prompts and quite directive,
(08:01):
giving it examples, so zero shotor one shot or many shot prompting.
those are going to be the key, for othersas we all get used to this new tool.
But those are the things that arereally helping me the most right now.
Sharon (08:20):
Yeah, I think that you're echoing something I've heard from
other guests, which is that right nowit feels like it's really great as an
assistant for you when you know how todo something and you can give it clear
directions and you can check its work.
that's something that it can do well interms of being able to speed up your own
(08:42):
process or help you with some parts ofit that are already pretty well dialed.
Or even as a thought partner, one thingthat I haven't tried, but now I'm curious
to try would be instead of giving it thedata set and asking for it to analyze
it, I'm curious if I could give it thedata set and say, what are the types of
(09:02):
analysis I could do on this data set?
like what kinds of, questions couldI answer or, so it'd be fun to cause
I do the same sort of thing in mywriting, where sometimes I'll say,
okay, this, these are some ideas I have.
Could you offer me five differentmetaphors that I might be able
(09:24):
to use to take this idea further,and use it as a thought partner.
So it'd be interesting to get it to just.
Remind you of Oh yeah, have youthought about presenting it?
It's this kind of graph, or haveyou thought about, looking at this?
because I have actually, I have put adata set into chat GPT and said, can
you tell me, some basic things aboutthe data that I knew I could, check
(09:50):
the answers because I was like, will itbe able to pull this information out?
It was accurate, but I was asking itvery simple things in terms of like,
how many people on this list are fromArkansas, or convert this to a table
where it shows what state, how manypeople per state, things like that.
So it could do some very basic things.
(10:10):
and that is something that I hope willbe possible in the future, is that
for people who just want somethingbasic, take my spreadsheet and turn
it into a table that I can put in myreport that shows the same information
and a slightly different visual.
It will make the barrier to entry for thatsmall amount of use of data better, so
(10:31):
more people feel more comfortable with it.
But right now we do have to be ableto check the work on everything.
Julie (10:39):
Everything.
Yeah, I hope we're there, too.
I really do.
And yeah, the same thing, I tellpeople, treat it like a summer intern.
They're really raring to go.
They want to please you.
They do not want to admitthey're making mistakes.
And you have to check everything.
But they're very nice andthey're very, yes, exactly.
(11:03):
But, but just with that, though.
Just with what you andI have talked about,
the possibilities, the next year,I just, I can't even imagine
what's coming down the pike for us.
So that's why I remain optimistic,cautiously optimistic, because we
(11:24):
haven't even touched on all theethics and the bias around all
of this.
but cautiously optimistic, withwhere this is headed and how it
can help us in our day to day.
Sharon (11:37):
Yeah.
I think the huge amount of variety ofways that could be used in the future.
One thing that's exciting, and Iguess I'll save this and then pivot
it to maybe a final question for you,which is I have been really enjoying
learning about ways that people arecombining technologies right now.
They're calling it multi modal.
(11:58):
So for instance, a, robot math tutorwho's able to have the student hold up
what they wrote on the paper and theycan actually, Take in that image and
know that the student wrote two plustwo equals five and then know that's
not correct and be able to, prompt back.
(12:21):
So being able to combine multipletechnologies, Is something that
just infinitely increases whatthese tools are capable of.
I'm curious What's something that you'refollowing closely or really loving
learning about right now that's on theedge of, what you've been hearing about
that you want to learn more about?
Julie (12:42):
Definitely the multimodal is on my radar.
I was at a, at conference lastmonth down at MIT about AI.
and it was all over.
It was all over the place.
So it's an area I have not been paying awhole lot of attention to prior to that.
now I'm thinking more.
(13:05):
And it's just, you think ofthese as a language model.
And so you think language, written.
But you're not thinking aboutvisual, oral, all those others.
And it's just.
All of a sudden, it's like you're walkinginto a completely new room in your house.
(13:26):
it's I had no idea this was here.
This is cool.
I always wanted a room like this.
so multimodal is definitely one of thethings that I'm keeping my eyes on.
And then what I was talking about earlier,the low code, no code solutions, there's
a lot more of those that are coming.
(13:46):
coming across my radar, I think thatif we're going to be using these tools
in anything beyond something like writemy 500 word essay or, something like
that, I think those are going to beone of the ways for people to really
(14:09):
be able to reap some serious benefits.
And then I'm always keeping my eye onwhat's going on with the data analysis.
It's just not there yet.
It's just not.
but I think it will be.
So I'm very eager to get some,some good evidence of that.
Sharon (14:29):
I heard someone lamenting the fact that they had found that
the tools were only as intelligentas a human three year old.
And I was like, wait,I've met three year olds.
Some of them are very intelligent.
that seems like we should bethinking, Wow, they're already as
intelligent as a three year oldand we're just getting started.
Julie (14:50):
No, I'm with you on that.
Sharon (14:52):
This is your host Sharon Tewkesbury bloom.
And I just want to pause brieflyto say that we really appreciate
that you've been listening to AIfor helpers and change makers.
I am supporting organizations who aretrying to figure out how to utilize AI.
Through my work at bloom facilitation.
So, if you are interested in howto bring new practices into your
(15:17):
workplace, I hope you'll reach outto me through bloom facilitation.com.
Now for more.
With my conversation with Julie.
many listeners, feel like AI is brand new.
They're racing to catch up.
They feel like they justheard about these tools.
but I'm a history majorby original discipline.
And so I love Figuringout, how did this evolve?
(15:39):
Where did this come from?
What were the steps that came before this?
And so I've also learned a bitabout the different waves of
technology advancement in this area.
And like you said, machinelearning is one of those.
so can you give us maybe a little bit ofbackground about what you had studied.
Julie (15:59):
Sure, it's interesting you mentioned you were a history major.
I was a French literature and historydouble major, and look at where we are
now, doing all of this techie stuff.
my head is in numbers all day,every day, and I love it, machine
learning has been around for a while.
it's a way of telling machines what to do.
(16:24):
similar to something likecomputer programming.
If you think back to the early days ofcomputer programming, you would, have
punch cards that would get fed in andwhatever you told it to do was output.
Machine learning was a step further it wasmore descriptive, more prescriptive to the
(16:46):
machine about what you wanted it to do.
And so the models were.
What I got in when I was in my doctoralstudies was with the quantitative
analysis part of it, which is heavilyinfluenced by statistics and econometrics.
So yeah, not everyone's goingto have any idea what that is.
(17:09):
It's basically using computers,using machines to help find patterns.
in a very high level way what machinelearning is about a part, under the
umbrella of artificial intelligence,that is one very well defined field
(17:35):
or subdiscipline that feeds rightinto everything else that's going
on today with those generative AI.
Because what is generative AI?
It's actually generating.
So traditional machine learning, typicalmachine learning is, it's not generating
(17:56):
anything based on what I tell it to do.
the generative part of it.
is like we see with chat GPT,with Claude, with some of these,
the language models, right?
It's actually generating paragraphs,output, conversations, that sort of thing.
So you can get into that muchmore fluid back and forth.
(18:20):
That's not typical machine learning.
how does it generate all this?
It goes back to prediction.
these large language models hooveredup all the text from the internet,
running equations and algorithmsto figure out what words were
(18:42):
most likely to beassociated with other words.
And so once you know that, then youcan start predicting what the next
word is going to be and So it's usingthat high level kind of thinking of
like machine learning But it's puttingit to use with predicting words and
(19:04):
the there's more to it But that'sbasically how they're tied together.
They build on top of the other.
Sharon (19:14):
that was really eyeopening to me.
I just completed this MITexecutive education course around
AI, and I didn't realize thatso much of this is statistics.
So much of it is based on probabilitiesand on being able to notice and
learn from patterns, then recognizepatterns, anticipate patterns.
(19:38):
I was not very good atstatistics, I'll admit.
I was really good at algebra.
I was really good at other kinds of maths.
Statistics was the bane ofmy existence in high school.
to know that these tools I'm excitedabout are based in statistics is amazing.
Julie (19:54):
Don't give up hope.
My first time with statistics in gradschool was nothing to write home about.
it wasn't, absolutely.
It's a different way of thinking.
Sure, it's math.
I'm good at algebra.
This is just a different way of thinking,just like calculus is a different way,
(20:15):
and you got to wrap your brain aroundthat, and it doesn't always happen
neatly within the confines of onesemester but I would encourage you, or
anyone else out there who's curious,keep at it, and you'll be surprised
by how you're able to grasp things.
Thank you.
Much more than you thought.
(20:36):
That's what I tell my students, too.
And it works.
It works.
They do get it.
But, yeah, it's all probability.
These generative AI, large languagemodels, it's all probability.
And that's wonderful, but that'salso something that we need
to keep in mind is a downside.
And it can be dangerous.
(20:56):
Because the same sorts of things that,The same caveats that I have to talk
about when I do forecasting and thosesorts of things for my clients and,
in my own work, I only have the pastdata, the history, to base it on.
And the same thing withthese large language models.
(21:19):
They're not necessarily built.
or good yet at conceptualizingsomething that hasn't happened.
And so they also can fall into thetrap of almost seeming like they
have blinders on because it's simplythe data that was fed into them.
(21:41):
They're not able to think outsideof that, at least not yet.
And that's where, humans, we reallyhave a leg up on the machines,
Sharon (21:53):
And I think for something relatable for a lot of people in terms
of how these tools and algorithms wereoriginally brought into many people's
lives is something like Netflix andhow they recommend movies to you
and how, based on your own viewinghistory and maybe what you spend a lot
of time looking at in the previews,they start to recommend other movies.
(22:17):
And, I do feel like I fight withthe machine sometimes where I'm
like, Don't put me in a box.
that's not the only kind ofthing I contain multitudes.
I like sports documentaries and Ilike cheesy romances and I like,
Nazi Germany history movies, becausethat was my, studies growing up.
So I'm sure Netflix is very confusedby me, but it's also a great example
(22:40):
of how, the machines and computersare trying to make sense based
on the data they already have.
And they're trying to notice patternsand predict things, but they can only
work with the data they already have.
They don't know that I've gonethrough some metamorphosis and I'm
suddenly interested in somethingbrand new that I've never shown
(23:01):
any inkling towards before.
Julie (23:05):
exactly.
They're advanced pattern recognition,trend recognition tools, if
we think about it that way.
Think about on your YouTube Timelineand every so often you'll get something
that pops up and says, do you wantsomething completely different?
And that's exactly so YouTube, atleast this is what I'm assuming.
(23:31):
They're trying to be ableto pull us out of that.
What might be a trap almostnot a trap, a dead end, maybe.
Maybe.
Sharon (23:39):
Yeah, I joke that my dad has reached the end of YouTube on his
subspecialties because he has somevery narrow interests that I think
he has watched every single thing
Julie (23:50):
Yeah.
Sharon (23:50):
in that.
If it's about earth moving machinesand it's been recorded by a bearded
white man, he has watched it.
So, and he knows because I make funof him about that, but All right.
So you are the data Diva.
At some point, I'd love to hear ifyou gave yourself that title or if
someone else gave you that title.
(24:11):
And then I'd also love to hear howdata, what you think people who
don't study data all the time needto know about data to, live in this
world today with how it's beingshowing up in everything we're doing.
Julie (24:28):
Mhmm.
I'll take that first question first.
my husband came up with it andhe came up with it years ago.
back when, I was in oneof my full time jobs.
And, when I went out on my own,doing my own consulting, my analytics
consulting, at some point I justjokingly said that in a networking
(24:51):
group and people really liked it.
And I was like, Oh, Iguess my husband was right.
so it comes from my husband, my
better half, I call him, and whatdo people need to know about data?
um, one research study recentlyfound that 93 percent of Americans
had some sort of math anxiety.
(25:15):
And at least in the businesscircles, the business ecosystems
where I am right now, a lot of timesdata equates to dollars and cents.
Equates to something that my accountantdoes for me, or my bookkeeper, or my CFO.
(25:39):
They handle that so that Idon't need to look at it.
And, my question back usually issomething like, You realize that
mountain of data has gold in it.
And your accountant or your CFO, theymight be looking in certain places, right?
(26:04):
For certain things so thatthey are able to create your
profit and loss sheet, right?
Your quarterly reviews.
There's a whole lot more in that mountainof data that if you had a different tool,
like some statistics, you be able to minethat for gold and diamonds and emeralds.
(26:29):
And I'll usually say something like that.
And then I'll give a couple of examplesabout because a lot of people think
my business doesn't have any data.
And I have clients that all they hadwas their financials, I was able to
find underperforming profit, profitable,offerings, I was able to help them to
(26:54):
untangle, what was going on with customerretention and why having their sales
team just continually bringing in morenew customers wasn't moving any needles.
And so with the data and the kindsof tools that I use, these more
statistical tools, I'm able to saywhy these things are happening.
(27:18):
And once you know why they're happening,then you can start fixing them.
And ultimately, businesseswant to stay in business.
They want to be profitable.
They want to be sustainable.
And many businesses, business owners,want to be able to have a legacy.
(27:42):
And If you're just drowning in redink and you don't know what's actually
making money, what's actually drivingyour customers to your competitors
or to just leave, which marketingchannel is actually working, you're
never going to be able to achieve that.
Wow.
Sharon (28:07):
fighting the good fight on that.
it made me think of this businessthat I know about, where I actually
had one of my first jobs ever.
It's a locally owned, familyowned retail business.
They just celebrated 50 years in business.
The daughter was able to takeover ownership from her father.
(28:30):
So she's now looking to thefuture of the business, and she
cares a lot about sustainability.
And so not only is she interested incontinuing to run a profitable business,
but because they do have this stablefoundation, she's now thinking, how
could we do even more for our community?
(28:51):
How could we do more for theenvironment or have a lesser
impact on the environment?
And I'm curious beyond just the numbersand cents, what other ways might
there be gold in those data hills forsomeone like that who's looking at
that sort of, what do they call it?
Triple bottom line business wantto do good for customers, the
(29:17):
environment and their business.
Julie (29:19):
I'll go back to what I say a lot of times to the skeptics
who say, my bank balance is fine.
My employees are paid.
I don't have any fines from the IRS.
What do I need to worry about?
And yeah, you don't needto worry about anything.
(29:39):
if that's your goal, then you're right.
You're good.
If you want to actually do somethingand you want to, be a good citizen in
your community, you want to be a goodsteward of the environment, anything
like that, you're not going to be ableto if you don't have a solid foundation
(30:07):
of what your business is doing wrong,what you can improve on in those.
it's all well and good tosponsor, Arbor Day or something.
Is that really making a difference?
And if your business isn't makinga profit, are you going to be
(30:30):
able to, sponsor any of that?
If your employees aren't engaged,healthy, both mentally and
physically, fulfilled with their jobs.
Are they going to be willing to, evenif you pay them for the day, right?
Sharon (30:48):
Yeah, and maybe part of what I was thinking is.
If you don't know, for instance, yourenergy usage as a business and how that's
also related to your costs, like how muchare you paying for heating and cooling
and, that's often a great place to startwith something like improving your impact
(31:08):
on the environment because there couldbe that win of this could lower costs
as well as be able to use less energy.
But those are data.
those are sources of data of being ableto track what are your energy uses?
How does that relate to yourability to run your business?
(31:28):
Are you staying open later hoursbecause you want to run a promotion,
but then does that cost youmore in energy or, other things?
So it's just interesting allthe layers that I can see of
potential data inputs to that.
Julie (31:43):
Yeah.
And if you've also got a solidaim in mind, it makes it easier to
convince people to track your data.
So no one likes tracking.
no one likes tracking anything.
I don't like tracking, did Ibrush my teeth this morning?
(32:04):
Did I brush it after lunch?
Anything like that.
And I see so many in our space outthere that almost like admonishing.
business owners andbusiness professionals.
Like you need to be tracking this.
You need to be tracking that.
Coming up with these checklists, theselaundry lists of things to, to track.
(32:27):
And I'm like, yeah, sure.
We can track everything from here tothe moon, but does it make a difference?
And is it going to really, move the needleon the things that our business needs to?
If instead, like what you said, Sharon,We have a commitment to reducing our
energy usage, then boy, someone isgoing to be on top of that, figuring
(32:53):
out what it is on a day by day basis.
today when we're recording this,I'm in the Northeast and it's
very hot, unseasonably hot.
Energy usage is way high.
What if I'm a business owner?
If I'm a manufacturer, whatam I going to do to offset?
all this extra energy that Ineeded to expend or buy today.
(33:17):
I have no idea, again,if I'm not tracking.
So it's a nice way also to really makeit clear to the members of your team.
Like this isn't justputting numbers in boxes.
We're going to do something withthis and you've got a role to play.
Yeah.
Sharon (33:40):
three possible barriers or hesitations people might have that
there's math anxiety that keeps peoplefrom digging into their numbers.
Perhaps they might be skeptical aboutthe benefits they'll get from actually
collecting tracking and using the dataand then tracking itself can be a chore.
(34:02):
is there anything else that you've seen?
that kind of keeps people fromdiving into their data and getting
the most benefit out of it.
Julie (34:13):
People don't draw enough pictures.
and by that um, what's the linguafranca of business in the world,
not just the United States, whenwe think about numbers and data.
It's Microsoft Excel, GoogleSheets, something like that, right?
(34:34):
And what is that?
It's columns and columns of numbersand rows and rows of numbers.
And if you don't put the format inthen it's all those ugly, number
signs in there that some peoplehave no idea how to get rid of.
And it's just, it's overwhelming.
even me, I don't know about you, but evento me, some of these are like, I just
(34:57):
feel like I'm being drowned in numbers.
And I can't blame anyoneelse who's, doesn't have fun,
like me, looking at numbers.
I don't blame them for not being all gungho about digging in and what have you.
The one thing that, I use with myclients, I use with myself, I use with my
(35:22):
students is draw a picture, make a graph.
doing that, the human brain, just theway that the human brain is hardwired,
it understands and perceives visualstimuli much more quickly and in a lot
(35:42):
of ways much more completely than othertypes, than, looking at numbers on a
spreadsheet, or even hearing things.
and so sometimes, most of thetime, the first thing I do is
start making, drawing pictures.
I make bar charts, scatterplots, and line graphs.
(36:07):
Those are the three that researchhas shown the human brain
is, best at perceiving with avery high degree of accuracy.
And so bar charts are not sexy.
They're not necessarily fun, but if you'vegot like several hundred thousand rows
of data and we've all been there, you canuse a bar chart and it will start cutting
(36:31):
through all the noise and it'll startshowing you, Oh, that's what's going on.
And then you can go from there ormaking a line chart or something.
honestly, I really do think making anExcel makes it easy enough, to simply
insert and it's got, a couple ofpoint and clicks and it'll make them.
(36:54):
It's a completely separate topicif they're actually necessarily
any good, the defaults that arein there, and that sort of thing.
But at least it's a way to cutinto the data and start to see
what the heck is going on here.
Sharon (37:12):
Yeah.
I have two follow up points on that.
One is also, sometimes it's not until youtry to make that picture That it requires
you to really ask yourself, what questionam I trying to answer with this data?
(37:33):
Cause you are asked like, whatdo you want to see on those bars?
What is your X axis?
What is your Y axis?
What do you want to see compared to what?
and I know I recently was working withAmeriCorps programs on recruitment.
And one thing we often ask them is,where are you getting people from?
Where do people find out about you?
(37:54):
That's important to know, but a lotof people have stopped at that level.
So they'll just have maybe a piechart that shows, okay, we got
this many through Indeed, this manyfrom a referral from a member, this
many from a job fair, et cetera.
But then I've asked them, okay,that's one piece of data, but is
(38:15):
that really what you want to know?
Or can we go deeper and say, Whichof those referral sources actually
led to members who didn't just applybut actually became members and even
better yet, if we have the data, whichof those actually were successful as
(38:36):
members and made it through to the end.
oftentimes, No one's ever gonethrough that data with them and
said, we could go deeper on this.
We could find out some reallyinteresting answers to our questions.
Not just the first default, which iswhere did the applications come from?
Julie (38:54):
yeah, exactly what you were talking about.
Lead sources and conversion rates.
That kind of pipeline analysisis something that I do for some
of my clients, and it's eyeopening, absolutely eye opening.
When we look at, the sources thatmight have the largest volume.
(39:18):
are not necessarily the oneswhere we get the stellar whatever
kind of customer or something.
Yep.
And it's very easy to show thatthen with a, some sort of visual.
Yep.
Sharon (39:35):
Excellent.
Yeah.
I also had done a training back in theday called making friends with metrics.
Okay.
Because, as you've found also, peopledo have a lot of anxiety about numbers,
and I work with a lot of groups throughuniversities, through grant funding,
etc., and they're required to trackdata, and they're required to report
(39:56):
it to a funder or to a higher up, butthey're not using it for themselves.
They're afraid of it.
And trying to help them
be empowered to actually say,this is your data, you can use it.
to help you feel more confident indecision making or to understand
where to put your resources on timeor to advocate for more funding
(40:19):
or whatever it is that you need.
and one example I show because it's ofteneasy for people to picture is that I used
to run the litter collection program,adopt a street, adopt an avenue, and so
we would do a very rudimentary type of.
data collection, which is that we'dhave a little slip of paper on the
(40:40):
bucket that the volunteers tookout with them to pick up trash.
And they would just record how manyvolunteers for how many hours picked
up how much, how many bags of trash.
So it wasn't fancy.
We weren't using any tech.
It was just a pencil and a littlepiece of paper, but then I could
input that into my spreadsheet.
(41:03):
And then we did actuallyhave a fancy GIS map.
And we were able to start overlaying thetrash, volume data over the map of the
city and we could start to see where'sthe higher concentration of litter.
In the city, geographically.
(41:26):
And then that led us to be ableto make better decisions about if
we have a big cleanup day, likeon earth day, are we going to send
people to the most popular park?
That's also the one that never needsany trash picked up because everybody
loves that park and they pick up theirtrash and keep it clean to begin with.
Or are we going to send them, to maybea part of the neighborhood that has.
(41:49):
It's, a lot of busy commercialarea with dumpsters that get flown
open and, everything gets broughtby the wind down to one low point.
And so all of the trashgets collected in that area.
so you start to see those patternsthat show up and it's just so
visual because it's literallylitter and it's in your city.
(42:10):
it helps people understand, Oh yeah,that would be really nice to be able to
know, like, where's the litter and thenwe can start to investigate why is it
always showing up in the same places andthat metaphor is the same for business
or same for people's operations of like,Where are the things showing up all the
time and why are they always showing up?
Julie (42:31):
Exactly.
Sharon (42:32):
thank you so much for Lending some of your insights
and wisdom and knowledge with us.
How can people learn more aboutyour work and maybe even contact
you if they have their own data thatneeds a diva to take a look at it.
Julie (42:49):
You got it.
I have a website, www dot janalytics, all one word.com.
If you go there, youcan book a call with me.
You can download, I've got a couple of,Free resources there to help professionals
(43:11):
and business owners try to untangleand make some sense of their data.
I hang out over on LinkedIn, so definitelycome over there, find me, connect with me.
Would love to meet you over there.
and I host live streams.
On Friday afternoons where Italk about something to do with
(43:35):
using data and AI in business.
So it's usually three on Eastern time.
So noon Pacific, early Saturdaymorning, in Australia, New Zealand.
I've got a few people overthere who watch, so I would
love to see anyone over there.
(43:57):
And then, pretty much.
Sharon (44:00):
Awesome.
Yes, I've been able to tune intosome of those lives and they're very
insightful, so I'd highly recommend them.
You can watch them live, or Ibelieve they also show up if you
can see the recording afterwards.
Do you know if that's true?
Julie (44:15):
You can, you can see the recording over on YouTube channel
at nhdatadiva.
There you go.
Sharon (44:23):
Awesome.
Brian AI (44:24):
Thank you for joining us on this episode of AI for
Helpers and Changemakers.
For the show notes and moreinformation about working with
Sharon, visit bloomfacilitation.
com.
If you have a suggestion for whowe should interview, email us
at hello at bloomfacilitation.
com.
And finally, please share thisepisode with someone you think
would find it interesting.
(44:45):
Word of mouth is our best marketing.
Popular Podcasts
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Law & Order: Criminal Justice System - Season 1 & Season 2
Season Two Out Now! Law & Order: Criminal Justice System tells the real stories behind the landmark cases that have shaped how the most dangerous and influential criminals in America are prosecuted. In its second season, the series tackles the threat of terrorism in the United States. From the rise of extremist political groups in the 60s to domestic lone wolves in the modern day, we explore how organizations like the FBI and Joint Terrorism Take Force have evolved to fight back against a multitude of terrorist threats.