Here are a few key takeaways from our 2015 Data Science Report.
Download the full report for all our findings.
Real people, real results:
The crowd understands language and intent better than machines. Algorithms get smarter, faster with human scoring.
Test your algorithms:
CrowdFlower contributors will score your current model and test it against your existing one so you know where you're succeeding.
Set it and forget it:
Our quality controls, plus a 24/7 labor pool, means your search relevance jobs run quickly and accurately, no matter when you start them.
- Nearly 80% of data scientists we surveyed loved their jobs
- That said, there's a shortage of data scientists: a full 83% of respondents said there aren't enough data scientists to go around
- 60% of data scientists spend most of their time cleaning and labeling data. 57% said it was the least enjoyable thing they do
- SQL, Hadoop, and Python are the most important skills for current and aspiring data scientists