Behind the scenes: A day in the life of a data scientist

2 months ago 22
PR Distribution

Helping others usage information is "like giving them a superpower," says the elder information idiosyncratic astatine an ag-tech startup, Plenty.

Portrait of Dana Seidel astatine  work

Data Scientist Dana Seidel astatine work.

Image: Dana Seidel

Dana Seidel was "traipsing astir agrarian Alberta, pursuing herds of elk," trying to fig retired their question patterns, what they ate, what brought them backmost to the aforesaid spot, erstwhile she had an epiphany: Data could assistance reply these questions. 

SEE: Snowflake information warehouse platform: A cheat expanse (free PDF) (TechRepublic)

At the time, enrolled successful a master's programme astatine the University of Alberta, she was funny successful tracking the question of cervid and elk and different cardinal foragers. Seidel realized that she could usage her mathematics and ecology inheritance astatine Cornell University to assistance measure a exemplary that could reply these questions. She continued her studies, earning a Ph.D. astatine University of California  Berkeley related to carnal question and the dispersed of diseases—which she monitored, successful part, by collecting information from collars. Kind of similar a Fitbit, Seidel explained, "tracking wherever you spell passim the day," yielding GPS information points that could link to onshore data, specified arsenic outer images, offering a model into the question of this wildlife.

Seidel, 31, has since transitioned from academia to the startup world, moving arsenic the pb data scientist astatine Plenty, an indoor vertical farming company. Or arsenic she would telephone herself a "data idiosyncratic who is funny successful spatial-temporal clip bid data."

Seidel was calved successful Tennessee, but grew up successful Kansas. She's 31, which she said is "old" for the startup world. As idiosyncratic who spent her twenties "investing successful 1 vocation way and past switching over," she doesn't needfully person the aforesaid manufacture acquisition arsenic her colleagues. So portion she is grateful for her experience, a grade is not a necessity, she said.

"I'm not definite that my Ph.D. helps maine successful my existent job," she said. One country wherever it did assistance her, however, was by giving her entree to internships—at Google Maps, successful Quantitative Analysts and RStudio—where she gained acquisition successful bundle development.

"But I don't deliberation penning much papers astir anthrax and zebras truly convinced anybody that I was a information scientist," she said.

Seidel learned the programming connection R, which she loved, successful college, and successful her master's programme started gathering databases. She said she "generally taught myself alongside these courses to usage the tools." The biggest accomplishment of being a information idiosyncratic "may precise good conscionable beryllium knowing however to Google things," she said. "That's each coding truly is, originative problem-solving."

SEE: Job description: Chief information officer (TechRepublic Premium)

The tract of information subject is astir a decennary old, Seidel said—previously, it was statistics. "The thought of having idiosyncratic who has a statistic inheritance oregon understands inferential modeling oregon instrumentality learning has existed for a batch longer than we've called it a information scientist," she said, and a master's successful information subject didn't beryllium until the past twelvemonth of her Ph.D. 

Additionally, "data scientist" is precise broad. Among information scientists, galore antithetic jobs tin exist. "There are information scientists that absorption precise overmuch connected precocious analytics. Some information scientists lone bash natural connection processing," she said. And the enactment emcompasses galore divers skills, she said, including "project management skills, information skills, investigation skills, captious reasoning skills."

Seidel has mentored others funny successful getting into the field, starting with a play Women successful Machine Learning and Data Science java hr astatine Berkeley. The archetypal portion of advice? "I would archer them: 'You person skills,'" Seidel said. Many young students, particularly women, don't recognize however overmuch they already know. "I don't deliberation we pass often to ourselves successful a affirmative way, each of the things we cognize however to do, and however that mightiness translate," she said. 

For those funny successful transitioning from academia to industry, she besides advises getting acquisition successful bundle improvement and champion practices, which whitethorn person been missing from ceremonial education. "If you recognize things similar modular manufacture practices, similar mentation power and git and bash scripting a small spot truthful that you person immoderate of that language, immoderate of that knowledge, you tin beryllium a much effectual collaborator." Seidel besides recommends learning SQL—one of the easiest languages, successful her opinion—which she calls "the lingua franca of information analytics and information science. Even though I deliberation it's thing you tin perfectly larn connected the job, it's going to beryllium the main mode you entree information if you're moving successful an manufacture information subject team. They're going to person ample databases with information and you request a mode to pass that," she said. She besides recommends gathering skills, done things similar the 25-day Advent of Code, and different ways to show a cleanable coding style. "What takes a bully magnitude of legwork, and until you person your manufacture job, it's unpaid legwork, but it tin truly assistance marque you basal out," she said.

SEE: Top 5 things you request to cognize astir information science (TechRepublic)

On a emblematic greeting astatine her existent job, moving from home, Seidel is drinking java and answering Slack messages successful her location office/ quilting studio. She checks to spot if determination are questions astir the data, thing incorrect with the dashboard, oregon a question astir works health. Software engineers moving connected the information whitethorn besides person questions, she said. There's often a scrum gathering successful the morning, and they run with sprint teams (meeting each 2 weeks) and agile workflows.

"I person a beauteous unsocial presumption wherever I tin interval betwixt assorted information scrums we do, we person a workplace show scrum versus a cognition squad oregon a information infrastructure team," Seidel explained. "I tin decide: What americium I going to lend to successful this sprint?" Twice a week there's a enactment meeting, wherever she is connected the bundle and information leads, and she tin perceive successful connected what other is being worked on, and what's coming up ahead, which she said is 1 of the astir important meetings for her, since she tin perceive straight "when a alteration is happening connected the bundle broadside oregon there's a caller request coming retired of ops for a bundle oregon for bundle oregon for information that's coming."

In the afternoon, she has a bully artifact of improvement time, "to excavation into immoderate contented I'm moving connected that sprint," she said.

SEE: How to go a information scientist: A cheat sheet (TechRepublic)

Seidel manages the information warehouse and ensures information streams are "being surfaced to extremity users successful halfway information models." Last week, she worked connected the workplace show scrum, "validating measurements that are coming retired of the farm, reasoning up astir the caller measurements we request to beryllium collecting, and reasoning astir the measurements that we person successful our southbound San Francisco farm, measurements streaming successful from a mates of 1000 devices." She needs to guarantee close measurement streams, which travel from everything from the somesthesia to irrigation, to guarantee works health, and reply questions like: "Why did past week's arugula bash amended than this week's arugula?"

The superior task is to cognize if they're measuring the close thing, and to propulsion backmost and say, "Oh, OK, what is it that you privation that information to beryllium explaining? What is the question you're asking?" She needs to enactment a fewer steps ahead, she said, and ask: "What are each the caller information sources that I request to beryllium alert of that we request to beryllium supporting?"

The toughest portion of the job? "I truly hatred not having the answer. I hatred having to say, "No, we don't measurement that happening yet." Or, "We'll person that successful the adjacent sprint." Balancing giving radical the answers with giving them tools to entree the answers themselves is simply a regular challenge, she said, with the eventual extremity of making information accessible.

And saying, "Oh, yes, that information is determination and it's this elemental query," or, "Oh, person you seen this instrumentality I built a twelvemonth agone that tin lick this problem?" is truly gratifying. 

"Helping idiosyncratic larn however to inquire and reply questions from information is similar giving them a superpower," Seidel said.

Data, Analytics and AI Newsletter

Learn the latest quality and champion practices astir information science, large information analytics, and artificial intelligence. Delivered Mondays

Sign up today

Also see

Read Entire Article