Get Started

Data Science for Dummies

Data Basics Stéphanie Griffiths

"How can you work with data scientists? You never liked math!"

That was the typical comment from my friends when I joined Dataiku. I heard so many misconceptions about AI, data science, and data scientists that I took on the challenge to clarify this agitation. If you are working at the C-suite level or in business functions as an analyst (marketing, sales, finance, or sustainability), before you think about AI, here are my three tips to demystify data science — whatever your math level may be.

1. Think About Data Science as a Jazz Band

When you want to maximize data, think about a jazz band with four instruments that need to be harmonized: code, statistics, industry expertise, and socio-economic context. It is impossible to find an expert on all topics. The same goes for data scientists. 

As a business leader, you are the invisible band leader (yes, in jazz, you don’t have a conductor). The value you bring when coordinating the band's work does not come from knowing more than any individual; your impact results from orchestrating synergy and letting each member express his/her individuality and explore interconnections. 

Always remember, “Data doesn’t belong to the data team” and collaboration across data and business departments is the only way to achieve engaging and intelligible performance.

jazz band

Tip 1: Accept that collaboration and chaos are at the heart of data science. 

  • If you are leading a company, give easy access to data (Don’t let IT or anyone make it more difficult, please. This old-fashioned way of working kills your ROI).
  • Empower business users to grapple with data and data science.
  • Confront perspectives: Bring people with diverse backgrounds to the table. Think design, sustainability, legal, and more.

2. Data Scientists Are Not Minions

In most companies, when the marketing or finance departments need data, they submit their (often urgent) request to data scientists. From their perspective, magic ensues, and they get an answer. End of story. Behind their screen, they might picture their teams like a minion — working day and night without question or rest. 

minions

Like anyone else, even if you like stats, math, or coding, you need meaning. Burnout and churn are relatively high within analytics/data science departments. If the onus is only placed on efficiency and automatization, tasks become repetitive and meaningless. Business teams should inverse the process and challenge data scientists on how they could generate more value and what means value for them within their company context.

Tip  2: Let data scientists play with data.

  • Ask them to come up with questions. It isn’t a one-way street.
  • Focusing only on efficiency won't get you there. 
  • Keep and grow your data scientist teams by giving them room for exploration, meaning, and creativity.
As the CDO of a top insurance firm told me, “I think about my data scientists as data artists.”

3. Think About Data Science as a System, Not as a Standalone Topic

Working correctly with data and data scientists is comparable to the circular economy, for optimal value:

  • Rethink: Approach each project with scalability in mind. By planning for deployment and constant improvement, agility will be de facto embedded. Conceive each project as a building block for another. 
  • Reuse: Think about the data, code, and narrative you could reuse.
  • Revive: Don’t start an algorithm from scratch; use coding skills to tailor an existing one.
recycling symbol

Think like the grandfather of physics and energy:

Nothing gets lost, nothing is created, everything transforms.” -Antoine Lavoisier, chemical scientist from the 18th century

Tip 3: Data science should be considered a system to scale projects from the start and constantly improve them.

  • Rethink, reuse, and revive your data projects. Don’t build once and throw it away. 
  • Keep code, narrative, and datasets to ensure you constantly improve and pass on knowledge to enable collective intelligence

Dataiku was created by data scientists to overcome tensions between art & science, creativity & control, one single good idea & systematic data processes. In this way, they opened the door for business users to contribute to this collective and creative process, reaping more value from data. The unleashing of data science creativity is what I most enjoy at Dataiku — if only my math teacher had gotten the memo!

You May Also Like

5 Ways to Get the Most Out of Your Local Data Science Community

Read More

The Ultimate Guide to Data Pipelines

Read More