Way too many people looking to get into data try running before they can walk: they want to train a neural network, for example, before they can append files together in Excel. But if you’re looking to get a first job in data and not just be a Kaggle celebrity (maybe?), chances are those Excel skills are going to be far more helpful than mere AI sophistry.
I’ve put some Google Trends searches together here to show how important these fundamentals really are… and that they seem to be getting more popular, not less, as the data industry matures.
Feel free to leave your own interpretation of these or other Trends reports in the comments!
Methodology
For each of these I am searching worldwide between July 1, 2017 and June 30, 2022 and embedding a tweet to the report. Where possible, I try to use the Google Trends topic rather than search term, except for where they aren’t available for all relevant terms to compare.
“Everyone’s becoming a data scientist!”
It’s easy to believe that everyone going into data is becoming a data scientist, and that this is the gold standard by which all other data jobs are to be measured. And how could that not be true? It is the sexiest job title of the 21st Century, after all.
The reality is that “data scientist” is just one of many data roles that an organization needs to be successful with data. In fact, it’s not even the first role that is needed.
As I explain in this article, hiring data scientists alone doesn’t make a data culture. There’s a whole hierarchy of needs and infrastructure that comes first. And if you come in hoping to train neural networks on data that barely hangs by the gridlines of its spreadsheet, you’re going to be miserable:
I think enough newcomers in data have heard this horror story by now. They see just how hard it is to get a “real” data science job and how many of them are a smokescreen for “glorified Excel analyst.”
Not that there’s anything wrong with being an Excel analyst! In fact, this puts you on the front lines of an organization’s data culture. You form the base of the pyramid for the data scientists to come in later. And with some direct experience as that foundation of the data pyramid, you’re in a much better place for an upward move to other data roles.
All that to say, not everyone is becoming a data scientist… it looks like data analyst is gaining some ground.
“Excel is dead, everyone’s using Python!”
Let’s stick to the discussion about being a “glorified Excel analyst” and ramp it up to the claim that “real data professionals never use Excel.” They only use serious tools like Python, right?
Yeah, well: message from planet Earth. Businesses use Excel. A lot. And anything that prevalent must fulfill some concrete need. In the case of spreadsheet data, it’s cheap, user-friendly analysis. Nothing wrong with that.
Python is great too, and it’s worth checking out at some point in your learning path. But again, starting with the foundations, I’d suggest you dig into Excel first. It’s much likelier that your boss and team will be using it versus Python, R or other coding languages.
I mean, look at this — search interest for Microsoft Excel as a whole is higher than all of Python… and remember that Python is used for many more purposes than data analysis (but so does Excel, to be fair…).
“Excel is dead, everyone’s using Alteryx!”
Now, onto the expensive business intelligence and data transformation tools (Alteryx, DOMO et al). Do you need to learn one of those to become a “hardcore analyst?” According to the data, I’m not so sure:
Again, stick with the fundamentals. Learn Excel. Take a look at Power Query, which is a data transformation tool built right into Excel. Just because a tool is hundreds of dollars more expensive doesn’t mean it’s better.
“I want to learn machine learning!”
This goes back to the data scientist fixation found among too many data newbies. Most employers are just not ready to benefit from data science and machine learning.
What they do need is help so that business partners can monitor data about what’s going on, and use that data to make more informed decisions. That is called business intelligence, and it’s beginning to surpass interest in machine learning:
Machine learning is really cool and it has revolutionized our lives. You should understand the basics if not just for the sake of being a more informed citizen. But for getting your first data job, it actually isn’t that relevant.
“Statistics don’t matter”
“Artificial intelligence” may be the buzziest of buzzwords. Certainly we can just sprinkle some AI on everything to make it modern, right? We don’t need to bother about fusty old statistics?
In fact, the search trends appear to differ:
Maybe the volume is just from a bunch of college students looking for help to pass their outdated coursework, but it appears that statistics still matters. Even in something so foundational to data as building a dashboard there is interest right now to add AI into the mix, but I’d encourage you to be wary of going all in on AI until you understand basic statistical concepts like sample size, the margin of error and so forth.
“But I want to be cutting edge!”
Look, it’s great to be an industry watcher in the data space. But you have to remember that unless you’re scoring a job in high tech, most organizations have a long way to go before adopting the latest and greatest.
What does this mean for your job search?
- Learn how to get the most out of Excel before diving into Python coding… or Alteryx no-coding.
- The fanciest algorithm is no help to an organization that just needs timely, relevant facts about the business.
- Artificial intelligence holds a lot of promise, but not that much for data professionals who aren’t even statistically literate.
This may be a bummer if you were expecting to be the next Andrew Ng to your thirty-person industrial supply company in suburban St. Louis. But spend some time with the non-data geeks on that team. They are in the majority, after all.
They are burdened with work. Budgets are tight. Like you, they’re doing their best. Unlike you, they’d rather do something else with their careers than learn about data. How can you make their jobs suck a little less by using data, and by consequence these concepts?
That tends to be through implementing the foundations, like cleaning spreadsheet data and using basic statistical processes to analyze it. The cutting edge stuff will come, eventually. Until then, enjoy the ride aboard your cushy job with fun dashboards (pun intended).
Leave a Reply