Explore

Analytics & machine learning

Data Scientist

Most of the job is cleaning data nobody documented.

Typical pay

$85k-$170k

How you get in

Quantitative degree plus a portfolio; master's is common

Outlook

Solid at senior level; saturated at entry

Data scientists build models and analyses to answer business questions and drive decisions. The romantic version is machine learning; the real version is a great deal of data cleaning, stakeholder management and explaining uncertainty.

A day in the life

  1. 9:00aCheck yesterday's model run and data pipeline
  2. 10:00aClean and join a messy new dataset
  3. 12:30pLunch
  4. 1:30pFeature engineering and model iteration
  5. 3:30pPresent findings to stakeholders who wanted a simpler answer
  6. 5:00pDocument assumptions so the result is reproducible

How to get in

Quantitative degree plus portfolio

Duration

4 years

Cost

$40,000-$180,000

Credential

BS in statistics, math, computer science or economics

  1. Complete a quantitative degree with real statistics coursework
  2. Learn Python, SQL and the standard data stack
  3. Build end-to-end projects on messy public data, not clean tutorials
  4. Enter as an analyst and move into data science internally

Master's degree

Duration

1-2 years after a bachelor's

Cost

$25,000-$90,000

Credential

MS in data science, statistics or analytics

  1. Hold a quantitative bachelor's
  2. Complete a master's with a capstone on real data
  3. Intern during the program — this is what converts to offers
  4. The most common entry credential for the title today

Analyst into data science

Duration

2-4 years

Cost

$0-$5,000

Credential

None — internal track record

  1. Start as a data analyst, which is far easier to enter
  2. Learn statistics and machine learning properly on the side
  3. Take on predictive work at your current company
  4. Move into the data science title with domain knowledge others lack

What people love

  • · High pay and intellectually demanding
  • · Remote-friendly
  • · Applies to nearly every industry
  • · Genuine influence on decisions

What wears people down

  • · Cleaning data is most of the work
  • · Entry level is very crowded and credential-heavy
  • · Results are often ignored for political reasons
  • · Title means wildly different things at different companies

From people doing it

Add yours