Recommendations For Noob In Data Degree

It's at least a month, 3 months in my industry, to start a new job. I'm not suggesting you lie about anything you're not over 100% certain you can learn. But then I'm capable of teaching myself anything really fast so maybe I'm only speaking for myself.

Just curious, what kind of skill that you can hide from the interviewer and not get caught, realistically?

Again I reiterate to everyone do not lie on your resume or during an interview. We will know, we will not tell you and during our get together after the interview we'll all agree that you are a very strong no. If a skill comes up that you haven't acquired, just say so and that your willing to learn. This is an acceptable answer because things change so fast in IT that we are all constantly learning new skills. Demonstrating the ability to learn new skills weights stronger then knowing existing ones.

One of my teachers has already prepped us on lying and has basically said the same (all of the teachers actively work in the field). My thing is, besides saying your Google and chatgpt skills are good, what do you say?! I'm sure more will come to me with more time in school.

Well, recently I was inteviewed by a dozen of people.

While not a job interview, it was an interview to get a spot in a department. At worse, I would land in a department I wouldnt really like to work with.

My interviews were mostly telling my work carreer, telling an overall about what I worked with in the past, why I left there and was here, finally what I wanted in my future carreer.

I managed to get the department I was aiming for, so I suppose I was successful so I would advise the same

It really depends on the position and how much you bothered to read up on it. Your resume should have a list of skills you have acquired and the job description should also have a list of "skills" that the hiring manage wrote in. Doing basic research into each of those will give you an idea of what they are. We've seen your resume, we know if a skill is or isn't listed, if it's listed we're gonna expect you to be able to back it up, if it's not listed then we may inquire and see if your able to communicate learning ability.

That is the entire point of those questions I posted earlier, they probe someone's ability to learn and adapt.

The only morally flexible advice I give is when people that I know are smart are applying for jobs, I tell them they should say they know SQL if the job asks for it even if they don't currently know it. If you get the interview, you can learn 80% of SQL in a single day of work before the interview.

Hmm I'd be careful with that, if knowing SQL is important and the interviewers have any experience with interviewing, their going to ask you to give a situation or describe your previous experience with it. Inexperienced interviews will ask trivia questions, experienced interviewers will ask candidates to describe previous assignments and work projects. Memorizing questions is easy enough, fabricating entire work projects is another matter.

Better idea is after going over a positions requirements, spend a weekend setting up a home project and mess around with it. SQL Express is free and will be sufficient for the ask. Write something silly like storing your bookmarks into tables and doing some sort of search and reference with them. Now someone can truthfully write they have experience with that technology.

If they want someone that know SQL, they also will want you to know NoSQL database, Datalake, Big Data, Spark Apache.

Things i never cared to know more than its basic description.

Where is Looker used in all of this? I keep hearing from people that's an important one to know.

I did a quick research, is Looker an alternative to Power BI?

I personally never heard of that, but mostly because I think Power BI is a major player in the BI stuff.

Power BI is a tool that basically does all data mining from you: you connect data sources, and Power BI will extract, process that data, and create dashboards for the CEOs to make analysis from it.

Our entire data section is going nuts over MS Fabric, Data lake and Power BI. It's been used to replace one of our major value added services, KRTA. It's were take we financial and performance data from most of the power companies in the US, do a ridiculous amount of data analysis to create a product that allows our customers to benchmark themselves against the industry. Previously it was done through a huge number of excel files and then published through QlikView onto our customer web portal. After it was converted to PowerBI it's become must to generate and we can publish it quarterly now instead of yearly.

They are now in the process of building out the products that allow our see US wide power meter trends. We have access to something like ~70% of all power utilization data in the US and it takes an entire department to make useful reports from it. Fabric and PowerBI looks to be capable of making much easier and letting us publish custom reports directly to our customers as an on-demand service.

Personally, if I am interviewing, I wouldn't ask questions about a skill that can be learn in a single day of work. And if I do, it wouldn't be a deciding factor to get picked anyways. So there are no point to lie about it, because it adds zero bonus points. Only negative point on a person's integrity.

To be honest, if someone asked me "do you know how to use Photoshop?" in an interview, I'd think this interviewer is either a total laymen or intentionally insulting people's intelligence lol. This kind of question is a waste of time.

The kind of question that I would ask is often more like "why/how do you do this" "how would you improve this" "how long does it take for you to do this". Then the quality of the answer to above questions decide whether you get picked or not.

There are three situations where SQL shows up on a job requirement list:
1. A DBA. I hope you already know SQL.
2. A technical job where you are also expected to know an actual programming language. If you are qualified for it and you somehow don't know SQL, you could learn it in a day and claim you used it in school but are a bit rusty.
3. A mostly nontechnical job (like a product manager) where they want to be able to say they are doing "data centric design" or whatever the other buzzwords are. They basically want managers to be able to query the DB and find out how the product is being used. The reality here is that DBs are uninterpretable without a guide anyway, and it is quicker for your engineers to write you a query you can run to pull stats than it is for them to explain how everything works anyway. If the person can fake it through these interviews, which would require a day of prep still, then they will never actually need to know SQL.

Anyway, it doesn't really matter because SQL is trivial to hit 80% competence on and the last 20% takes forever and depends on the implementation but unless you are a DBA they probably won't expect you to know it.

Anyway, to the OP, "data science" is as broad a field as the name implies. Some people are PowerBI monkeys, some are screwing around in R still, some (particularly in finance) are "modeling" in Excel, but most are using python.

I recommend getting at least competent at coding.

I said in an interview once I could teach CFD (Computational Fluid Dynamics, basically gas and liquids simulation) if I needed to. I haven't done much of it before and I've never taught it, but I knew that if I had to I could. It would only have formed a small part of a module and not the entire purpose of the module.

I've also said I can teach Fusion 360, which I never have but I have expertise in several extremely similar programs.

I had never used Rhino when I first agreed to teach it, 2 weeks later I was teaching advanced surfacing to 2nd year students. I had never used Blender, done CG, key frame animations, or anything to do with AR when I said I'll teach myself so I can teach it. That took about a month. If it works for an employer to employ me knowing I can't do or haven't taught something before, in the belief that I had the knowledge and skills to pick it up, then I don't see why the reverse enrages people.

There's probably a degree of bias (as the employer/interviewer, you want the ability to decide whether the candidate is capable of learning rather than them deciding). But, there's also a large potential consequence to lying, especially in tight industries/fields. It may not be the best idea to burn bridges suggesting you have skills you don't and coming off as a liar/fool. If you're intelligent enough to learn the skill rapidly, presumably you're also intelligent enough to demonstrate that to the interviewer.

Thank you!!

Another question: Can you please explain the difference between Data Science and Data Analytics in industry? (How they are the same / different.)

I was speaking to a classmate a few days ago who said they want to go for data analytics. But I don’t understand the difference because it seems like there is a ton of overlap on skills / requirements.

If you already have previous teaching experience, you probably can convince your employer because teaching as a skill is transferable between different different subjects.

For example, if I am hiring a music teacher, and the applicant can play piano and violin equally well, but they only have experience teaching piano for a few years, not violin. I can believe that they should have no problem teaching violin too.

Maybe there are some key difference between teaching piano and violin, but the difference shouldn't be that huge especially if your students aren't at high level.

The same can be said for teaching how to use different softwares in university too. Learning how to use different software doesn't take long, it generally took me a few days to learn the important functions of softwares that I use irl. So I can see why an experienced teacher would have np convincing employers that they can teach different software even if they have never done it before.

Because the hard part here is "teaching", not "learning software". And you already proved your "teaching skill" is good in the past.

In the industry is different though. There is a pretty huge gap between certain professional skills and it isn't easy to just say "I can do this" then somehow get through it with just words. Most of the important questions won't be "can you use blender?" level of question.

There is a lot of variability, but everywhere I have worked the Data Analyst job title basically means "person who makes plots but not models." It is purely a "less than" role relative to "data scientist" and probably caps your salary ceiling pretty low. I wouldn't train in school to be a data analyst, but if it ends up being your first job out of school because people don't want to hire you to be a data scientist it wouldn't be the end of the world.

BTW, after today Im done with data science at least until eventually being allocated to work directly with it.

Im also unenrolling on my Deep Learning course 3/5 done and doing an UX course instead.

At this very moment, I already think most ML stuff more annoying than interesting.

I think only Gen AI seems interesting to me because I actually like drawing and AI helping creative work is a very nice field, but all the rest seems underwhelming to actually work with it.

Looking back in perspective, I dont think that programming is a core skill of data science per se. I dont classify running scripts and using libraries as some noble role for programming to have. Sure, you sure need some computer skills to create the interface of models with users, but thats pretty decoupled. Maybe you work in very low level and can make those algorithmns run better, but obviously the established libraries are the place to go for that.

Tuning a model relies more on heuristics and statistic. And statistic is surely the worst subject I had in my life. The only one I failed twice before graduating from it.

Obviously there's a huge bias in employment if you don't value yourself highly. At the time I was in the favourable position. "Lying" was a bit of an exaggeration to what I meant anyway, "exaggerating" was probably more appropriate.

UX design will all be 90% automated with AI in a few years and the bottom end of UX designers completely wiped out. Not a hugely safe bet either IMO.

I dont think so tbf.

UX is a heavily empathy based work. There is no way AI to understand well user needs.

Plus, UX is huge understimated task, and the one directly involved with the end user.

Maybe stupid people with narrow view of the process wont hire a focused UX professional, but im completely sure an engineer or developer that have this skill has his/her own weight in gold

Is modeling just = to ‘machine learning’ algorithms (such as regression analysis?).

Because if so, this seems like a pretty easy jump from Dat analysis -> science.
The only skill difference I can see would be math and coding.

Quite a lot of people are advocating for AI generated personas and even synthetic users (AI 'people'). I think a lot will go this way.

I think the main thing AI is going to teach us is that people aren't really that complex. We like to think we are, and that that's what makes us human and differentiates us from animals. We're all unique snowflakes in our minds, but really we're all driven by the same things that push for us to use one app or service over another.

The methods you should use depends on the problem. "ML" is probably overused as a buzzword, although it will help in interviews, and my guess is that GLM is still the true workhorse driving industrial data science.

It is true that every data scientist needs to be able to do data analysis, but not every data analyst can make models. If you are going to get certified by a university as a data scientist or analyst, you should get certified as a scientist because that is probably the easiest way to verify your capability. If you have a degree that says Data Analyst, you have to convince someone to take a risk on you when hiring for a data scientist and that is going to depend on the labor market conditions.

https://www.nngroup.com/articles/synthetic-users/


Well yeah, imo, AI is probably good enough for businesses with a low bar. Or no budget.

In today's business though, low bar isn't good enough to win the competition. If you are a business owner who settle for low bar, you will lose in the market.

So people with high bar will still use real human until AI is 100 times smarter than they are now. Which aren't going to happen any time soon due to hardware processing power limitation.

Just FYI, I've been trying to implement AI to work since...May 2022 too? Then everytime when AI has a "breakthrough" I tried to implement them to workflow too.

I also followed many so called "AI guru"'s blog, looking for new ways to implement their discovery to work.

But the result has been disappointing so far. The quality of work produced by AI just isn't anywhere close to commercially competitive quality. Those "AI guru"'s discovery on the internet don't solve quality problem either. They only generate clicks for outsiders because it looks cool.

Trust me, I WISH I can click a button and have AI do all the work so I can sleep all day, but it is not happening atm.


After lots of effort now I can delegate maybe 3%-5% of daily easy task to AI. Yay!

So technically, I can be one of the 50% cool people that claimed to use AI at work.

But this AI is far, FAR from replacing me atm lol. I just don't see it happening anytime soon.

The bar there is not actually "as an expert, do I *** that I do a much better job than AI" but instead is "in the opinion of my boss's boss's boss, do I think two guys using AI would do a comparable enough job to this department that I should fire the department and cut costs?"

That many steps up the chain, they won't even recognize the actual tradeoffs. It'll all be hubris, unearned confidence, and "looks the same to me"-level assessments with a healthy dose of hype train. They may also end up trying to hire you back in a month when the product lights on fire.

That said, I have seen no evidence this is actually happening anywhere that wasn't already on fire and trying to cut costs.