As with any industry, not all data science career prospects are made equal. Hundreds of data science position descriptions that I’ve viewed have had huge variations. There are a number of causes behind this, but one of the main ones is that many businesses do not understand what a data scientist performs or the advantages of using data science effectively.
There are companies out there that say they’re looking for a data scientist when they really need a data analyst. Some people think a data scientist is the same as a data architect or data engineer, while others disagree. Creating a data model for a database is what the company I worked for paid me to do. These are not the kinds of information models you need. The reply I gave was, “(With proper credit to Obi-wan Kenobi)” – I make no claims to being a data architect.
Therefore, if you’re in an interviewing situation, you’ll need to have a good eye for true data science jobs as opposed to postings that claim, “someone told us we needed a data scientist, so we put together this description.” There are four things to keep in mind whether you’re doing a phone interview or a face-to-face meeting.
Question 1: Why do you believe you require a data scientist?
It’s important to ask this initially, but it’s not the most pressing concern. You should get to the bottom of why the company thinks it needs a data scientist by talking to the CEO, CFO, CTO, or whoever is doing the recruiting. They commonly lump together or confuse several data talents into one massive term: data scientist, even if they may need some of the abilities that a data scientist brings to the table, such as data processing and visualization.
Data scientists seldom create data warehouses or applications that rely on relational databases. Not only that, but data scientists are not programmers in the traditional sense. In contrast, a data scientist’s job description will increasingly call for them to access data stored in data lakes, data registries, and data silos using SQL and other query languages.
Analysis often necessitates data staging from these disparate sources.
Therefore, a data scientist is probably not needed if the recruiting team just needs someone to create and construct a data strategy. That is, until recently. Although some data scientists may be capable of doing so, they are not seeking someone to develop a relational database application. A data scientist isn’t what’s needed if a company just needs someone to collect information from various sources, clean it up, and organize it into a usable format for use in a business intelligence tool. These are the abilities of a business intelligence developer. Although this is a talent that is shared by many data scientists, it is not essential to the role of a data scientist and is seldom used.
Question 2: What business problems do you want the data scientist to address?
This is an important follow-up inquiry because of the first. The company probably knows it faces business issues, but doesn’t know how a data scientist can help. Make sure they can explain in detail how these problems are related to data science. They probably don’t need a full-fledged data scientist, but rather only a few specific skillsets. A third party may have suggested they hire a data scientist. There’s also the possibility that they saw it in a magazine. However, they should be able to explain why a data scientist is needed rather than, example, a data analyst.
It’s not a terrible thing in and of itself that I’ve been employed to do visuals; I do have the ability to make visualizations. I can use R, Qlik Sense, Tableau, and even Excel, but so can a data analyst, a business analyst, or a fresh college graduate. Furthermore, they are not as demanding of payment as I am.
A data scientist is unnecessary if a company has no plans to do inferential, predictive, or prescriptive analytics in the near future.
Question 3: Are you in possession of a data warehouse?
Quite a few businesses are floundering because they don’t have a solid understanding of their data. A data scientist might be hampered by the immaturity of data. They may be inundated with data from several sources, much of it is scattered over numerous MS Access databases, Excel spreadsheets, and other document formats. Part of the information might be kept in a local database or in the cloud. Having big data or a large data set does not guarantee that a company is mature enough to sustain the job of a data scientist.
A data scientist will have a hard time if there isn’t a data warehouse or at least a plan for bringing together all of this raw information for usage by the various divisions of the company.
Can you picture yourself explaining to a client that you have the skills necessary to complete the assignment, but that it would take you a year to collect and organize the data? That quickly, even before you’ve had a chance to evaluate it? That’s an unacceptable delay for any company, but it’s not impossible. A data scientist may need to consult with a SME more than once during the course of their work.
I’ve worked in environments where I couldn’t talk to subject matter experts. The word “overwhelmed” accurately describes their feelings. I wasn’t able to finish what I had started. In other words, it’s not enough to just have data; you need to have accurate data. Once, I waited nearly a year for a better perspective on a 15-minute sequel enquiry. There was a backlog of work for the subject matter experts, and I was unable to gain access to the view to address the issue. My productivity plummeted. You don’t want to be working in an atmosphere like that.
Question 4: When do you expect to see results?
The significance of this inquiry cannot be overstated. In what time frame do you expect them to see the results? Data science, like so many other things in life, is a slow process. Even if everything falls into place, a data scientist will still need to spend a lot of time manipulating, analyzing, and visualizing information. After that, it’s time to go back and do some more data processing, visualization, and analysis, after first validating the work with relevant stakeholders. The change will not happen all at once but rather gradually. It lacks glitz and glamour, too. Don’t go into data science thinking it’s all glitz and flash.
Because of this, you should probably go elsewhere if the company you’re interested in working for has an unrealistic expectation of outcomes in 30, 60, or 90 days. To put it another way, you’ll set the stage for your own failure.
As a result, make sure your inquiries are relevant and get honest replies. To compare what they’re telling you with what you know, take notes and give them a quick tour of your idea of what a data science project should look like. When you do this, and only then, will you be able to determine the best culture for efficient data science.