Data Science is mainly a Human Science
Data Science is mainly a Human Science: in latest KDnuggets Poll, 62% of data scientists worked only/mostly on human data. At the same time, most data scientists worked with both human and machine data.
By Gregory Piatetsky, @kdnuggets, Sep 23, 2014. (re-posted with permission of http://www.kdnuggets.com)
The results of the latest KDnuggets Poll:
What kind of data science did you do in the past 3 years?
show that Data Science is mainly a Human Science, with 62% of respondents working ONLY or MOSTLY on data science for humans. Only 14% have worked only or mostly on machine data. The “Other” votes could represent people working on non-human biological data or purely scientific data, like astronomy.
At the same time, about 60% of respondents worked with both human and machine data, casting doubt on the assertion that there are 2 types of data scientists: ones that work with machine data, and ones that work with human data.
“What kind of data science did you do in the past 3 years: [195 votes total]
>Only Data Science for Humans (63) 32%
>Mostly Data Science for Humans (59) 30%
>About equally Data Science for Humans and for Machines (40) 21%
>Mostly Data Science for Machines (17) 8.7%
>Only Data Science for Machines (11) 5.6%
>Other (5) 2.6%
The table below gives a regional breakdown for 4 regions with sufficient representation, with 5 “Other” votes excluded.
We note that Europe and Asia have higher machine-data experience than the US, while in Latin America almost all data scientists who responded work with human data.
| Region | Only or Mostly Human Data | Equally Human and Machine Data | Only or Mostly Machine Data |
|---|---|---|---|
| US/Canada (51%) | 66% | 19% | 11% |
| Europe (27%) | 49% | 30% | 21% |
| Asia (11%) | 52% | 24% | 24% |
| Latin America (8.2%) | 94% | 0% | 6% |
| ALL | 62% | 21% | 15% |