Not many people can say they have been looking at the same math problem for a month.
Welcome to the world of the data science generalist. Complicated problems with complicated solutions that need even more complicated explanations. If using breaking it down, sorting it all, and delivering solutions in iterations sounds thrilling, this is the job for you.
People with instincts and skills to engage with and solve problems has been booming. Thanks to the data revolution, companies need more data dealt with than they even know.
We’ll cover the day-to-day specifics and ups and downs of this career thoroughly.
Data Science Generalist
As the title suggests, the job is a subset of data science. Data science specialties have a place and build on the same skillset as a generalist with more, well, specialties.
To better understand the scope of the field for data science careers, we’ll hit the broad strokes first. Then we’ll explain the day-to-day intricacies.
No job is without its problems, so we’ll also cover that. Finally, we’ll talk about where the field is heading.
A data science generalist’s job is to practically create their job. Companies know they need to work with data, and they need to extract relevancy from it. They don’t always know how to collect that data, how to interpret that data, or how to ask the right questions of that data.
Your job will include a grab bag of skills including the ability to collect data through logistic and linear regression. You will need to build and control and teach algorithms.
To gather data, you will need CSV and relational database knowledge. To process the data, you need to understand imputation and deduping techniques.
You will also need to have a firm – but not total – grasp of how to visualize data with programs like Pandas and Matplotlib. To construct and refine algorithms, you will need to know both your Python and Matlab.
You do all of this while also collaborating with other team members (if you have them). Then you will need to bridge the gap between production and management.
Nuts and Bolts
A typical day in the life of a generalist involves finding a problem to solve or continuing to find a solution to the problem you’ve had in front of you for a month.
It takes a certain business sense to identify if the problem matters. Not all problems actually affect real people. You can find the most elegant solution to an issue only to realize that nobody actually suffers from the issue, to begin with.
So you will take your discovered problem and look for data that both support the existence of the problem and offers a window into solutions.
Then you start plugging in the good stuff. Often data will exist, and if not, you will have to find it. Finding data is rare, but parsing through data for something you can use is far more common.
You will create an algorithm to sift the mountain of useless data for the stuff you want. The algorithm will pull out some aberrations and some junk, and you will have to revise it. In a larger company this process is aided or under the purview of a machine learning engineer.
Eventually, you will like the data you have and can move forward.
From here it is all about how to present the data to anyone who didn’t spend a month looking at it and needs to make a decision. This is where the creative ability to visualize data comes through.
Often this is a team effort, but in a small company, you may be putting it all together yourself. You present your findings and then move on to the next issue or a revision of the current one.
One of the more difficult issues you will face in data science careers is fighting the expectations of executives. They will under and overestimate the value of what you can do at the most inopportune times.
While you know that data can be used for public good and corporate good, you will also have to face the public’s reluctance.
Even if you understand a problem, you can’t ensure that your solution will be accepted. When a group, be it your own team, the higher-ups, or the public needs convincing, it is important to know how to convince someone.
Facts and numbers don’t communicate well. They do to you, but that is your whole job. However, other people might be swayed by emotion and persuasive pandering. Be prepared to deal with some aspects of these issues.
Fortunately, the future of data science generalists is bright and wide. Large companies need do-it-all middle personnel that can reinforce and emphasize teams. Small companies need a cheap hire that can do the work of a team in a pinch.
Being a generalist also gives you the ability to start on the ground floor of an industry and specialize over time. This provides both short and long-term job security.
More companies, those that are not data companies, awaken to the importance of data collection and management every day.
Problems tend to magnify as the world keeps spinning, so those don’t seem to be in short supply either.
Open Your World
Even as the title of the job shifts to fit people’s expectations, a data science generalist will keep doing the same work. That is to say, doing a lot of different things in a broad heading that people don’t have the time to parse out.
Still, if you didn’t like to parse words, mince details, and come out technically correct, you probably wouldn’t be in this field.
To find more career opportunities in your chosen field look to our dedicated list to help you find data science jobs.