149: I Asked DeepSeek How to Become a Data Analyst (It scared me)
I tested DeepSeek-- an emerging AI platform that makes ChatGPT look ancient! I asked it to outline a comprehensive roadmap for becoming a data analyst. What it said scared me (Spoiler: it basically copied my SPN Method)!
Listen to NEXT: My interview with StatQuest!
https://www.youtube.com/watch?v=nqtQUg4mZ9I
💌 Join 10k+ aspiring data analysts & get my tips in your inbox weekly 👉 https://www.datacareerjumpstart.com/newsletter
🆘 Feeling stuck in your data journey? Come to my next free "How to Land Your First Data Job" training 👉 https://www.datacareerjumpstart.com/training
👩💻 Want to land a data job in less than 90 days? 👉 https://www.datacareerjumpstart.com/daa
👔 Ace The Interview with Confidence 👉 https://www.datacareerjumpstart.com/interviewsimulator
⌚ TIMESTAMPS
00:00 - Introduction
01:05 - Skills
01:27 - Do you need a degree? DeepSeek answers
01:59 - Projects and portfolio
02:43 - Networking and job search strategies
04:55 - Interview preparation
10:15 - FindADataJob.com and PremiumDataJobs.com
11:30 - InterviewSimulator.io
🔗 CONNECT WITH AVERY
🎥 YouTube Channel: https://www.youtube.com/@averysmith
🤝 LinkedIn: https://www.linkedin.com/in/averyjsmith/
📸 Instagram: https://instagram.com/datacareerjumpstart
🎵 TikTok: https://www.tiktok.com/@verydata
💻 Website: https://www.datacareerjumpstart.com/
Mentioned in this episode:
📰 Join 18k Other Analysts in My Newsletter
Get 1 weekly newsletter helping you take the next steps in your data career! Check it out: https://datacareerjumpstart.com/subscribe
✨ Try Julius!
This episode is brought to you by Julius – your AI data analyst companion. Connect to your database and/or business tools, pull insights in minutes–no coding required. Thanks, Julius, for sponsoring this episode. Try Julius at https://landadatajob.com/Julius-DCP
https://landadatajob.com/Julius-DCP
1:1 Coaching w/ Avery
I've opened 12 slots on my calendar to do 1:1 coaching calls. What we can do: fix your resume (it may suck), audit your LinkedIn (also may suck), mock interview you (prepare to sweat), or tell you exactly what to do next This is a test. Never done it. Might never do it again. Limited spots. Book now.
Transcript
If you haven't heard of deep seek yet, just wait.
2
:It is a new AI platform that supposedly
is maybe even better than open AI
3
:and chat GPT reasons being number one
is currently free and open source,
4
:meaning it is a lot more easily
accessible to the entire world.
5
:And number two, it has some
pretty cool thinking properties
6
:where it actually shows you it's.
7
:Thought process, and you can see how
the computer algorithm and the AI is
8
:actually thinking through the problem.
9
:So I thought, why not ask
DeepSeek on how to become a data
10
:analyst and see what it tells me.
11
:So let's go.
12
:So I'm on DeepSeek's website and
I'm going to turn on the DeepThink
13
:R1 model, which basically is their
better reasoning model right here.
14
:And I'm going to ask, I
want to be a data analyst.
15
:What should I do?
16
:And hit enter.
17
:One thing about DeepSeek is it
is a lot slower because it is
18
:actually going through and thinking
through your process step by step.
19
:So although OpenAI and ChatGPT may be
faster, DeepSeek is actually taking
20
:its time to think through the problem.
21
:And you see those thoughts right here.
22
:So it says, okay, the user
wants to become a data analyst.
23
:Let me start breaking
down what that entails.
24
:First, I need to outline the
essential skills they need.
25
:Data analysis requires technical skills
like SQL, Excel, maybe Python or R.
26
:I already love that the maybe is here.
27
:If you guys have watched any of
my videos or listened to any of my
28
:podcast episodes, you know that I
am not a fan of learning Python.
29
:First, and I think SQL and Excel
are a better place to start.
30
:Visualization tools like Tableau
or Power BI are important too.
31
:Yes, they should also understand
statistics and data cleaning.
32
:Next education.
33
:Do they need a degree, a
background in math stats or CS
34
:helps, but it's not mandatory.
35
:So far, I really agree with
what DeepSake is saying.
36
:Obviously it helps to have those.
37
:STEM background, some sort
of math or stats or computer
38
:science, but it's not mandatory.
39
:Maybe mention online courses and
certificates as alternatives.
40
:Platforms like Coursera,
edX, Datacamp come to mind.
41
:Okay.
42
:Certifications from Google or
Microsoft could also add credibility.
43
:I'm going to be honest.
44
:I don't think either of those
add a ton of credibility.
45
:The Google and the Microsoft certificate
certificates still don't mean anything
46
:in the data analytics world, but
they don't hurt your credibility.
47
:So I guess that's fine.
48
:Projects are crucial for experience.
49
:Oh my gosh.
50
:I agree with this.
51
:That's part of the SPN method, right?
52
:In order to land the data job,
you have to have the skills.
53
:You have to build projects and
you have to have a network.
54
:So projects are crucial for experiences.
55
:They should work on real world
data sets to build a portfolio.
56
:I totally agree with that.
57
:Where can they find data sets?
58
:Kaggle, government portals,
maybe APIs, suggest projects like
59
:sales analysis or COVID trends.
60
:Totally.
61
:We do sales and marketing analysis
in the accelerator program.
62
:Uh, and we do some healthcare analytics.
63
:We don't do COVID because COVID has
been done a bajillion times, but
64
:we do actually diabetes building
a GitHub portfolio and personal
65
:website would showcase their work.
66
:Uh, yes, I would reverse those
where I would do a personal website
67
:first versus a GitHub portfolio.
68
:If you're just trying to land the
data analyst job, if you're going for
69
:a data scientist or data engineer,
that's a little bit different, but
70
:I liked that it didn't just say Gab.
71
:I liked that it said personal website,
networking and job search strategies.
72
:Yes.
73
:That's the end part of the SPN method.
74
:I like DeepSeek so far.
75
:LinkedIn is important, it sure is, as
well as attending meetups or conferences,
76
:tailoring resumes to highlight relevant
skills and projects, applying to
77
:entry level positions like junior data
analyst or business analyst, maybe
78
:internships if they're just starting out.
79
:Continuous learning is key in this field.
80
:That is so true.
81
:Encourage joining communities,
following blogs, podcasts.
82
:Hey, shout out to, uh, podcasts.
83
:If you want to follow a podcast,
check out the Data Career Podcast.
84
:Staying updated with trends
like AI and machine learning.
85
:Also, soft skills like
communication and problem solving
86
:are often overlooked, but vital.
87
:So far, so good.
88
:I really agree with
what DeepSeek is saying.
89
:Wait, did I miss anything?
90
:Okay.
91
:It's so weird to have.
92
:A, an AI model say something like that.
93
:Oh, maybe the tools in more detail.
94
:Excel for basic SQL for databases.
95
:Python slash R for advanced analysis.
96
:Visualization tools like Tableau,
Power BI, maybe even Looker.
97
:Cloud platforms are becoming more
relevant too, like AWS and Google Cloud.
98
:Also the importance of understanding
the industry they want to work in.
99
:Man, 100 percent I agree with that.
100
:Domain knowledge can set them apart.
101
:Yes, it can.
102
:Like if you're trying to break into data,
whatever previous experience, whatever
103
:previous job, whatever previous education
you've had, that's, what's like going
104
:to be your in, into the world because
data analysis is in every industry.
105
:So it's like, that is
where you need to focus on.
106
:Uh, so I liked that.
107
:I got that.
108
:Domain knowledge can set them apart.
109
:Maybe healthcare, finance,
e commerce, et cetera.
110
:Tailoring projects to specific
industries might help.
111
:Once again, a hundred percent correct.
112
:This is why in the accelerator.
113
:We do nine different projects
in nine different industries.
114
:So no matter what industry you're
interested in, we have a project that you
115
:can pitch to a hiring manager recruiter.
116
:Hey, look, I have, I've done something
very similar to what your job requirement
117
:is asking you to do, asking me to do.
118
:Here's my evidence that I can do this.
119
:Should I mention freelancing or
volunteer work as a way to get experience
120
:that could be useful for building a
resume, maybe suggest platforms like
121
:Upwork or volunteering for nonprofits.
122
:I love the volunteering for nonprofits.
123
:I think that's great.
124
:I think most people do.
125
:Who try to go the freelancing route.
126
:It's pretty difficult because
freelancing is basically, you have
127
:to be good at whatever you're doing.
128
:So in this case, data analysis, but
you have to be good at data as well.
129
:I mean, you're just starting out
to be perfectly Frank, unless you
130
:have freelancing experience some
other way, you're bad at both.
131
:So it's not my favorite place.
132
:To start if I was you, but it didn't, it
just said, maybe we can consider them.
133
:What about interview preparation?
134
:They might need to practice SQL queries,
case studies, and presenting their
135
:projects, behavioral interviews as well.
136
:Resources like sleek code for
SQL practice or mock interviews.
137
:Let me structure this step by step
start with learning technical skills,
138
:then education and certificates,
building projects, networking, job
139
:applications, and continuous learning.
140
:Make sure each section is clear
and actionable avoid jargon.
141
:So it's accessible.
142
:Maybe include resources
and examples for each step.
143
:Becoming data analyst involves building
technical skills, getting practical
144
:experience and showcasing your abilities.
145
:Here's a step by step guide
to help you get started.
146
:All right.
147
:Just like the thinking was very good.
148
:And obviously that's not something
you get with open AI and chat GPT.
149
:So I loved hearing how it was thinking.
150
:And that way I can kind of
interpret the context of the
151
:results with what it was thinking.
152
:Uh, let's go ahead and take a look
at what it officially suggests.
153
:Build foundational skills, technical
skills, Excel, master pivot tables,
154
:VLOOKUP and data visualization.
155
:That is super simple.
156
:I like that a lot.
157
:SQL.
158
:Learn to query databases.
159
:Join aggregations.
160
:Resources.
161
:Mode.
162
:Analytics.
163
:SQL tutorial.
164
:Kaggle.
165
:Learn SQL.
166
:Both two great resources for learning SQL.
167
:Python slash R.
168
:Focus on libraries like pandas,
NumPy, and visualization tools.
169
:Resources.
170
:Data camp.
171
:Free code camp.
172
:NumPy.
173
:I mean, it's a, it's fine to learn pandas
is the big one in visualization tools.
174
:NumPy actually pandas is built on NumPy.
175
:So by knowing pandas, you kind of can
do a lot of the stuff that NumPy can.
176
:There are some things that you need to
use NumPy for alone, but there's not
177
:that many data visualization, learn
tools like Tableau, Power BI or Looker
178
:Tableau, public Tableau training.
179
:Great.
180
:Statistics, grasp basics like hypothesis,
testing distributions, and regression
181
:Khan Academy, StatQuest YouTube channel.
182
:Shout out StatQuest and Josh Starmer.
183
:I've done a video with Josh in the past.
184
:Uh, we'll have that in
the show notes down below.
185
:I also love that it did statistics
last out of all of these, because
186
:often people are like, don't you have
to know statistics to become data
187
:analyst and you do, but it's maybe
less important than you realize.
188
:For instance, you can do a VLOOKUP
really without doing any sort of.
189
:Tons of statistics.
190
:There are aggregations there like sums
and averages, mins and maxes, but the
191
:chances are you've probably learned that
in high school or college at some point.
192
:So something more complicated like
hypothesis testing or regression,
193
:those are important too, but not as
important as knowing how to do a VLOOKUP.
194
:I love that it's including
soft skills here too.
195
:Soft skills, storytelling,
present insights clearly
196
:to non technical audiences.
197
:Problem solving breakdown, complex
questions into data driven solutions.
198
:A lot of the times, and honestly, like if
you go look at the platforms, it suggests
199
:like data camp and free code camp,
they really miss on the storytelling.
200
:And it's really important because
we had cool math, like on the
201
:podcast recently, and she's like,
people don't take you seriously.
202
:If you're unable to communicate
your story and your work.
203
:Effectively.
204
:So it's often overlooked.
205
:One of the things we do inside the
accelerator is one, we post our stories
206
:on LinkedIn in our about section.
207
:And we try to post on LinkedIn
often, but also we post our projects
208
:on there with a written writeup.
209
:Um, and that written form of
communication, we actually will teach you
210
:how to do video communication as well.
211
:Like if you like recorded yourself with
a Google slide presentation or with
212
:your dashboard, but those types of.
213
:Communication and storytelling
opportunities are really what can set
214
:you apart as an aspiring data analyst.
215
:Next section, formal
education and certificates.
216
:So degrees optional.
217
:Love that it says optional.
218
:That is very friendly.
219
:I appreciate that.
220
:I'd agree in math, statistics, CS, or
economics is helpful, but not mandatory.
221
:I agree.
222
:Online certificates, Google data
analytics certificate, and the
223
:Microsoft data analyst associate
power BI plus Excel and the IBM data
224
:analyst professional certificate.
225
:Obviously those are all good and fine.
226
:Gain practical experience.
227
:I love that it's saying this.
228
:This is awesome.
229
:Projects, analyze data from Kaggle
UCI learning machine learning
230
:repository or government portals, data.
231
:gov.
232
:I think that is a hundred percent correct.
233
:Example projects, sales trend analysis,
COVID 19 impact dashboard, customer
234
:segmentation, obviously not a whole
lot of like, this is just an intro.
235
:Right.
236
:So it's not really giving us any more
information on these, but that's great.
237
:Portfolio share code on GitHub and
create visualizations in Tableau public.
238
:Build a personal website using GitHub
pages or WordPress to showcase projects.
239
:Love that it got GitHub pages.
240
:That is something that
a lot of people miss.
241
:Now I need to make a different
video on this, but GitHub pages
242
:is very separate from GitHub.
243
:It is like it's from GitHub, but
it's like a separate product.
244
:It's free, but it's basically like the
ability to build personal websites.
245
:And I love that it got it.
246
:I personally recommend card now.
247
:We'll have a link in the show notes.
248
:To to check out card thing is
the best and easiest place to
249
:start building your portfolio.
250
:Freelance slash volunteer offer
services on Upwork or nonprofits.
251
:I like helping nonprofits more
because I think they could offer more
252
:support and like a more formal role.
253
:Once again, I think freelancing
on Upwork, especially if you've
254
:never freelanced before, it's not
going to lead very far because.
255
:Freelancing requires a ton
of business experience.
256
:You have to know how to market yourself.
257
:You have to know how to
ask a lot of questions.
258
:There's no one checking your work.
259
:So I would lean on the volunteer
side versus the freelance, but
260
:I don't mind them mentioning it.
261
:Okay.
262
:Number four, network and apply for jobs.
263
:LinkedIn optimize your profile
with keywords like data analysts
264
:and connect with professionals.
265
:You guys, I can't tell you how
important this first line is.
266
:And it really, if you just read it,
you're like, okay, that makes sense.
267
:What does that actually mean?
268
:You guys, this is one thing we
talk about in the accelerator.
269
:The more you put the term data analyst
on your LinkedIn profile and your
270
:resume, the better you'll be off.
271
:ATS is the LinkedIn
recruiting algorithm is dumb.
272
:One of the ways it actually like checks
to see how relevant you are to, for
273
:instance, if you're applying to a data
analyst role is how many times do they
274
:have the word data analyst on their.
275
:LinkedIn page.
276
:And that phrase can be anywhere that
could be in your headline that can
277
:be in your about section that can
be in your experience section that
278
:can be in your education section.
279
:For instance, if you just put aspiring
data analyst in your experience section,
280
:that actually almost works as good to a
computer as putting the term data analyst.
281
:So that is really key job platforms,
entry level roles, junior data analyst,
282
:business analyst, reporting analyst.
283
:Those are all goods search on
LinkedIn indeed, or specialized
284
:sites like well found.
285
:Yeah.
286
:Well found angels lists.
287
:I'm a fan of, but not really
for entry level roles.
288
:They're more senior roles there.
289
:Instead, I would try
something like findadatajob.
290
:com or premiumdatajobs.
291
:com.
292
:Those are two job boards that I run where
we try to be more entry level friendly.
293
:Meetups, attend events, data science
meetup, Pi data or virtual webinars.
294
:I think that's great.
295
:That's a form of networking
and obviously a great option.
296
:Okay.
297
:Number five, ACE interviews, technical
prep, practice SQL on leak code or hacker
298
:rank review, statistical concepts and
case studies, behavioral questions.
299
:Use the star method to answer
questions about teamwork and problem
300
:solving and portfolio walkthrough.
301
:Be ready to explain your projects,
goals, process, and impact.
302
:Once again.
303
:This feels really good because most
people are all about the technical prep
304
:and the technical prep is important,
but I would say, honestly, at least half
305
:of my students who land jobs through
the accelerator program, never really
306
:even have a formal technical interview.
307
:The other 50 percent definitely do.
308
:And it's good to be prepared using
things like leak code or hacker.
309
:I prefer things like strata scratch,
data lemur or analyst builder.
310
:Instead of these, they're just
more data oriented instead of like.
311
:Computer science and stuff like that, I
think, but I just want to point off that
312
:most people ignore behavioral questions.
313
:And that's one of the things I try not
to ignore with interview simulator.
314
:If you guys go to interview simulator.
315
:io, this is my interview
platform where you can practice
316
:your behavioral questions.
317
:And then I love that it has the portfolio
walkthrough as well and being able to
318
:talk about your projects because really,
if you can get an interview and you
319
:can say, Hey, I have this portfolio.
320
:I've done this project that's similar
to what I would be doing on the job.
321
:I think that is an opportunity for you to.
322
:Try to take the interview kind of
by the reins and flip it on them.
323
:And they ask you questions about
your project versus just like
324
:asking random statistical concepts.
325
:So that's going to make you
feel more comfortable and
326
:make you look better as well.
327
:Number six, keep learning, stay
updated, follow blogs like towards data
328
:science and podcasts like data skeptic.
329
:Those are both great.
330
:I would add data career podcast to the
podcast, but if you're listening to this,
331
:you're probably already following our
podcast, advanced skills, explore machine
332
:learning, scikit learn, cloud tools, AWS.
333
:Google Big Query or A B testing.
334
:I think those are, I mean, that's fine.
335
:You're always going to be learning in this
world, but it didn't really talk about job
336
:applications and applying like you don't
want to just like go to advanced skills
337
:without job hunting a ton because you can
get paid to learn machine learning and
338
:cloud tools and A B testing on the job.
339
:Most entry level roles, even maybe
middle roles don't even require that.
340
:Uh, example, learning path, Excel,
SQL, Tableau month, one to two months,
341
:three to four Python and statistics and
month five to six build three or four
342
:portfolio projects and apply for jobs.
343
:Now, if you've listened to any of my
episodes previously, you know, that I
344
:think that most people, um, if they're
willing to put in, you know, 10 to
345
:20 hours a week can cut this in half.
346
:And there's like certain things that we
can do where it's like, we're not going to
347
:spend an entire month learning statistics,
an entire month learning Python.
348
:You just don't need to, when you're
landing your first day at a job.
349
:And also like why wait till month
five and six to build your portfolio
350
:projects and apply for jobs.
351
:In the accelerator, you'll have a project
built within your first 10 days, your
352
:first project built 10 days guaranteed.
353
:Like if you just put it in the hours,
like done, we'll have your first project.
354
:Build and then we want to start
applying for jobs, you know, well,
355
:before we hit the six month mark,
we're probably talking to the two
356
:month mark, if I'm being honest,
because applying for jobs isn't art and
357
:you'll get better at it as you go on.
358
:But overall, I don't hate this
plan at all by combining skills,
359
:projects, and networking.
360
:Holy crap.
361
:Did they steal that from me?
362
:Skills, projects, and networking.
363
:That's the SPN method.
364
:I came up with the SPN method.
365
:I'm the only one who's ever put those
things right next to each other.
366
:Skills, projects, and networking.
367
:Call it the SPN method.
368
:I built it myself.
369
:I have a notebook somewhere over
here where like, I just, I wrote
370
:down my whole like framework and like
tried to figure out what to call it.
371
:And we ended up landing on SPN.
372
:Wow.
373
:You'll position yourself
strongly for a data analyst role.
374
:Stay smart, stay consistent
and iterate based on feedback.
375
:Overall, you guys, like, I feel
like this was a pretty good roadmap.
376
:Right here.
377
:Like I'm pretty impressed by this and, uh,
it's not the worst thing on planet earth.
378
:Like it, it did a better job
almost in the instructions.
379
:I think of like the Python where it was
like maybe Python and R somewhere up here.
380
:Right.
381
:And, but I think it did great on
mentioning the behavioral questions.
382
:I think it did really good on
the networking and the, and
383
:the projects and the portfolio.
384
:I thought I did great talking
about GitHub pages overall.
385
:I think if you followed
this plan, you would be.
386
:Pretty well off.
387
:I mean, this plan is basically what
I outlined in my previous episodes.
388
:It's basically following the SPN method.
389
:I mean, literally it says by
combining skills, projects, and
390
:networking, you'll position yourself
strongly for a data analyst role.
391
:And I agree like that, the SPN
method will set you up exactly.
392
:This way.
393
:So, uh, I really like this from deep seek.
394
:I'm going to play around with this more.
395
:If you guys want to follow the
SPN method, please consider
396
:joining the accelerator program.
397
:This is basically a coaching led and
group cohort learning style where
398
:you're basically going to do all
of these things, but we're going to
399
:give it to you exactly step by step.
400
:You're not going to have to go figure
out like, you know, how do I learn
401
:data visualization and Tableau public?
402
:Or like what courses should I take?
403
:We'll give you the exact roadmap.
404
:We'll teach you the exact projects.
405
:We'll give you the exact data to
build your projects, to learn the
406
:skills and to grow your network.
407
:We'll show you exactly
how to actually optimize.
408
:Like what does it actually mean
to optimize your profile with
409
:keywords like data analyst?
410
:So that's of interest to you.
411
:We'll have a link in the show notes
down below and let me know what
412
:you guys want me to do next with
deep seek down in the comments.
413
:Should I try to analyze data?
414
:Should we compare it to
something like chat GPT?
415
:Let me know in the comments down below.