17

I started working at my current company earlier this year, and I'm concerned that I am wildly unqualified for the role.

In the part of the world where I live and work, the standard practice is to apply to a company (not to a role). In the interviews the company and candidate "get to know" one another, and there will of course be discussion of what sorts of skills the candidate has, as well as what sorts of work are available at the company. But it's not at all strange for a candidate to not know what department they'll be working in before their first day. (Although generally placement will be made after training and evaluation in the first few weeks. I assume that this usually results in a fairly good match, since it is still so widespread).

I applied to my current company because of my interest in their field, and in the interviews I received a very favourable impression of their culture.

They seemed interested in my work history with data analysis. And they asked whether I knew any Python, to which I responded that I had studied a bit, privately and not in connection with my previous position.

I was glad to receive an offer, although it only enumerated where my position would fit into their hierarchy. I nevertheless thought I had a fairly decent idea what my work would entail, and this all seemed natural for this part of the world.

I discovered on day one that I was to be a data scientist. And, what is more, the only data scientist in the company. I am the entire department.

I am not a data scientist. I'm mathematically literate. I took calculus and physics in college. And I'm quite good with computers (as a consumer). I understand programming structures, and how computers access and use data. But all as a layman. I have a few years experience in data analysis, but this was a completely backwards company that had their entire data base saved in excel files with filenames incremented for each year-month, going back twenty years. And I've taken an introductory course on data science and programming in Python, SQL, and R, but this was just introductory.

The company seems to have understood when hiring me that I was not a data scientist, but they intended that I would become one. How they imagined this would happen is an open question, because there are no training programs or experienced colleagues to guide me, and there is the constant expectation of output, by means of actionable reports.

I've done my best. I've made a training plan, and studied the contents of Hadley Wickham's "R for Data Science" and "Advanced R" books. And I've kept up my output, turning in reports that (apparently) look like data science, but which amount to "here's a graph of sales by month, with a trend line. Here it is by product tag. Here it is by category."

Admittedly, the company has not done well with communicating what they want from any given analysis. It's instead been a continuous repeat of "here's a set of sales figures and dates. Do analysis." And despite my attempts to wring out what exactly their questions or interests in this data is, or to acquire more or more useful data, there doesn't yet seem to be much chance of improvement.

So far my work has achieved absolutely nothing. Nothing I've submitted has had any impact on the company's future direction or current actions. And although it can't be said that I've made no progress in developing my skills, this really isn't the way to go about training a functional data scientist. They've been wasting their money on me.

And yet they don't seem to realize it, or not to care. Both my immediate supervisor and the president of the company are very interested in what data science can offer, and very eager to know what sorts of things I'm working on, but they don't know nearly enough to accurately judge a data scientist's skills or output. And they're (so far) perfectly happy with what I've been outputting, despite it changing nothing.

I feel terrified, both morally and personally, that this is wrong, and that it's going to explode in my face before long. I'm not contributing, and so I'm getting paid as if under false pretenses. I'm not comfortable with this. And personally, I'm not getting any meaningful experience. I'm not meaningfully developing as a professional, and if they ever wake up and fire me, or I grow a spine and quit (as I fear, morally, I really ought to), then I'll have wasted my time completely, and be that much worse off trying to find another job.

I've told the company, three times now, to three different people (including HR and my immediate supervisor), that I'm not confident that I'm doing good work. That I'm not confident that I have the skills necessary for this position. And that the best piece of analysis that I can give them is that they should save my salary by getting rid of me, and then hiring an experienced actual data scientist. But they take the position that everything's fine, and they know it'll take time to get used to things, and all of that. But my sense continues to grow that they just really don't understand how untoward this situation is.

I don't know what to do. Both morally, as a person who want to do the right thing yet feels horribly in the wrong. And personally, as a person who has to keep making a living and having a career of some sort, and who might just be digging their own grave.


I know that other questions have been asked regarding feeling unqualified for a position, but I believe there are two main differences between this question and those.

・Firstly, the answers to those questions have often been "Actually, you're completely qualified! Pat yourself on the back and relax." I am completely confident that this is not the case in my situation. Perhaps I am mistaken, but this question will take it as a given that the employee in question (namely, me) is unqualified.

and

・Secondly, in this case the unqualified person makes up the entire department, and there is therefore no one in the company with the expertise to properly evaluate their skills.

4
  • 4
    "In the part of the world where I live and work, the standard practice is to apply to a company (not to a role).... it's not at all strange for a candidate to not know what department they'll be working in before their first day" — which part of the world is this? Oct 27 at 11:41
  • 2
    I don't really see what's offtopic here, anyone care to comment?
    – Muzer
    Oct 27 at 13:54
  • @Muzer: I'm not super-familiar with what's on- and off-topic here, but maybe the question lacks a clear goal we can address. Oct 27 at 14:24
  • Do you actually want to become a data scientist or not? Is this a direction that interests you? Oct 28 at 17:29
24

Sounds like you're a data scientist to me.

Remember, data scientists often spend much of their time cleaning data (even if it might not be the 80% figure commonly quoted). If your entire job involves cleaning data and producing visualizations, it sounds like you're doing data science to me. You might not be doing the most impactful data science in the world, but producing graphs of sales by product is still useful for management staff to make decisions with (even if those decisions might not be visible to you). For instance, it might help them decide if a product is selling sufficiently to justify its existence, or to make decisions about inventory levels or about when they need to expand manufacturing for particular products.

Of course, if you want to make higher-quality insights, you might need higher-quality data. You can't make observations about customer behavior without customer data, for instance. You may want to investigate the potential to gain access to that sort of data, if you haven't already done so, and don't already have access to it. It might be possible for you to draw connections between the items that customers often buy, classify customers into groups who buy certain products, or even derive predictions of which customers are pregnant, for instance. If you don't know how to do that, it might suggest a path for you to investigate.

Also, I'd recommend trying to deal with impostor syndrome, and certainly to stop complaining to HR about it! It's possible that if you keep on with it for long enough, they might believe you and you'll find yourself out of a job.

3
  • 9
    Data scientists often want to make real fancy data science. Customers (in this instance, the higher ups) are often happy to get anything out of data at all and even the most basic of methods are a big step up and they are happy to have someone who knows how to do this. I have seen more than one data scientist come out of university and face a reality clash how mundane his job is! Depends a bit on company of course. so +1 to the answer
    – Benjamin
    Oct 27 at 7:37
  • @Benjamin And this situation is certainly not restricted to data science!
    – anderas
    Oct 27 at 11:54
  • 1
    "If your entire job involves cleaning data and producing visualizations, it sounds like you're doing data science to me" - that's data engineering and data analytics/business intelligence, respectively. One certainly does those things as part of data science, but to say they're data science in themselves for that reason would be like saying anyone who writes any sort of code (for any purpose) is doing data science (because writing code is very often a part of data science). But then again there are varying opinions about what data science actually is. Oct 27 at 12:52
14

You're not doing anything morally wrong

I feel terrified, both morally and personally, that this is wrong, and that it's going to explode in my face before long

...

But my sense continues to grow that they just really don't understand how untoward this situation is.

I don't know what to do. Both morally, as a person who want to do the right thing yet feels horribly in the wrong. And personally, as a person who has to keep making a living and having a career of some sort, and who might just be digging their own grave.

I think you're catastrophizing the situation a bit. No-one's died, no-one's gone to prison. You've been very clear and honest with the company about your skills and experience. You're not misleading them, and you're putting in a lot of effort to complete their very vaguely-defined tasks, while trying to both define a new role at this company, and learn how to actually do it, almost entirely on your own.

It sounds like you've convinced yourself that the company isn't getting sufficient value out of you, but that's not your decision to make. You're not the CEO. It's up to the company how they spend their money, and right now they're happy spending it on your salary while you try to become a data scientist. I believe data scientists are currently in high demand, so they might not even be able to hire anyone with more experience than you.

But is this a job you want to learn?

personally, I'm not getting any meaningful experience. I'm not meaningfully developing as a professional

Brief counterpoint: learning how to train yourself, and establish an ill-defined role at a company where management themselves don't really know what the role is, is actually pretty meaningful experience, even though it's really tough. Situations like this aren't hugely uncommon. Most companies have to innovate how they do things, and if you can get good at doing that, you're incredibly valuable.

But I do think the most important question to ask yourself is whether you want to be a data scientist. Establishing a role at a company while learning how to do that role is, as discussed, really hard work.

If you're enjoying getting into the data science, great! This is an excellent opportunity for you to learn how to do the job, and a excellent opportunity for the company to get a (probably relatively cheap) data scientist who's learned how to do exactly what they need, and nothing else.

But if you really don't want to become a data scientist, trying to force yourself to become one (especially with so little support and guidance) is a really tough task. For your own happiness, it might be worth trying to move into a job you actually do want to do.

2
  • 7
    Hey maybe you could swap jobs with this guy. Oct 27 at 11:42
  • 1
    To me this is the answer that actually hits the point. It's hard to tell from OP's post but the real crux of it is, do you want to be doing this? If OP has no actual interest in data science then moving is only natural. If OP would like to learn then this is a great opportunity, as long as they are comfortable with the idea of independent learning, even if it takes a long time.
    – Muzer
    Oct 27 at 13:53
7

Take a deep breath. You're actually fine.

Your company hired you, knowing full well that they don't know anything, and that they like you, and like what you have done so far, and you seem to be able to build good rapport. This is great.

You also know that you're not a data scientist, which is, actually fine too. Here is a pyramid on what needs to be done (this is a multi-year project):

enter image description here

If your company has data in spreadsheets, you guys don't even have the first step of Data Acquisition done yet. You need to learn this pyramid, give your leaders an honest projection about what needs to happen before seeing results from the top tier.

This company may want you to lead them into the next digital age, so you may have to do some courses (on their dime of course...) The way I see it, they're not remotely setup to do any real "data science" work, and even hiring a real data scientists isn't going to solve it because none of the foundation or infrastructure even exists.

Go back to your boss, ask them if it is their long term goals to have real data science as part of their business, and plan your next move from there.

3
  • I'll be honest with you. "Real" data science is actually quite mundane. You're a company using algorithms, not researching them. A lot of times you plug in models, tune it, run it against existing historical data see if it makes sense, and then let the algorithm make projections. Your boss then greenlights some of them and you let them loose and see if you're right. You're not doing data science with a pile of spreadsheets. For one, that's not remotely enough data.
    – Nelson
    Oct 27 at 4:33
  • 3
    All this requires significant infrastructure (quick access to months to years of historical data), AND you need to be able to access relatively fresh data to confirm whether the modelling worked. This is all extremely heavy duty work and is going to require teams of people when up and running.
    – Nelson
    Oct 27 at 4:37
  • "ask them if it is their long term goals to have real data science as part of their business" - although one should ask them this in a way that someone who doesn't know what data science is would understand. Oct 27 at 13:04
2

It sounds like you're dealing with some major Imposter Syndrome.

Ultimately, if your employer is happy with what they're getting out of you, and you're being paid. That's your employee-obligations filled and you're golden.

That they're doing it by the least efficient or effective means possible is in no way your problem.
That they could do better with someone trained and qualified is not your problem either.

As you note though, if they "wake up" and realise that they could simply hire someone who actually knows the material from the get-go and get better results.
That'd probably put you out of a job and you'd have to find something new.

It's therefore in your interest to get good at the job and do so quickly.

What shouldn't be understated though is that you're in an amazing position here:
You have the resources to learn a whole field and currently have low expectations on your productivity, meaning you have the time and resources to push your career further.
Exploit that to the hilt.
Become the Data-Scientist you believe you should be.
It's always okay to make yourself overqualified for your job.
If your employer is presented with your additional capability, you will be in a position to ask for raises, and more than capable of going to another company if you don't get one.

Remember that Self-Taught is not a bad thing either.

You might be self-taught, but in my field as a software developer everyone is.

It is understood that a university course will only teach broad theory and introductory practical material. If you want to be a professional developer, you will need to learn multiple programming languages, plugins, tools and a massive spread of esoteric knowledge which changes and expands constantly from year to year.

No university or college could ever teach a person all the knowledge and skills required to do my job in a few short years.

Almost every software developer is therefore primarily self-taught, usually on the job, and we have to keep learning because the technology is constantly changing underfoot.

Being self-taught is the normal state of things in my field, the IT Industry and indeed Data-Science.
Don't let that worry you.

Build a spread of practical duties that you perform for your company, make notes, make a clear list of what you can do. Try and expand that list as you do more things.
Then when you eventually move on to another business, put that list in your CV. Remember that the difference between a professional and an amateur is mostly whether you get paid to do it.

If the imposter-syndrome flares up, read that list again.
Remember that every single thing on that list is something you can do that most people don't know how to do.
That exclusivity is why you are being paid, and the more things you can put on that list, the better you justify your paycheck.

2
  • "You have the resources to learn a whole field" — I mean, kind of. It's unclear whether the company is providing time and money for training alongside the duties of the job. Oct 27 at 14:18
  • Well yeah, there's that I guess. Oct 27 at 14:32
1

In today's world you can be the leader of a very large nation without being qualified. One would imagine the company knows your capabilities and knowingly gave you this role. If you are bothered by your performance you could have this conversation with your boss and proceed from there.

1
  • 1
    . . . you can be the leader of a very large nation without being qualified, isn't this the worst?
    – Arriel
    Oct 27 at 15:29