“white building with data has a better idea text signage” by Franki Chamaki on Unsplash

Are Data Scientists Happier than Other Developers?

Comparing data science developer habits and job satisfaction with the rest of the community based on the 2017 annual stack overflow survey.

Ioannis K Breier
Sep 5, 2018 · 5 min read

Introduction

Stack overflow is publishing an annual survey of the developer community covering a large array of questions and providing interesting insights into the life and work of developers worldwide.

The survey data covers 64,000 reviews from 213 countries and territories and there were more than 150 questions as a part of the survey.

I wanted to find out whether there are identifiable differences in the work habits and job satisfaction between developers that work in the data science / machine learning field compared to the rest of their colleagues.


Question 1: What is the proportion of developers working in the data science field.

The following developer types are indentified by the survey question:

Which of the following best describe you?

Respondents could choose more than one field, so there is quite some overlap between functions. I categorized as data science developers the ones that have included one of the following descriptions in the developer type field:

'Data scientist',
'Machine learning specialist',
'Developer with a statistics or mathematics background'

I will use the term data scientist and data science developer interchangeably referring to the above category of developers.

How many are they?

Fig 1: Counts and Percentages of Developer Types.

From a total number of 51,392 developers 6,353 have included a data science related developer field in their description which is 12 % of the total.


Question 2: Are their working habits different?

Where do they work from?

Let’s start by looking at their remote working habits as captured by the question,

How often do you work from home or remotely?

Fig 2: Percentages of Responses to the Remote Work question per Developer Type.

There seems to a clear difference in favor of data scientists across the answers with 38% of data science developers working from home or remotely a few days per month and only 26% responding never, compared to 35% and 33% for the other developers.

Do they program as a hobby as well? ?

Programming can be a lifestyle that extends well beyond working hours and this is captured by the following question,

Do you program as a hobby or contribute to open source projects?

Fig 3: Percentages of Responses to the Hobby Programmig question per Developer Type.

Here, we can see some obvious differences as well. The percentage that do not program outside work at all or only program as a hobby is higher for other developers at 19% and 49% compared to 17% and 32%. Data science developers on the other hand, have a higher percentage of contribution to open source projects (7%) and both (33%) compared to 6% and 26%.

In total, 83% of data science developers, program out of work compared to 80% of other developers.


Question 3: Are they happier?

And finally the question we have been waiting for, as captured by the job satisfaction field,

Job satisfaction rating

Fig 4: Job Satisfaction Rating per Developer Type

There is a clearly a higher percentage of data science developers that rate their job satisfaction at 8 and above and a higher number of other developers the rate their job satisfaction as 7 or lower.

The differences are not big but they are consistent. This trend is also reflected at the average job satisfaction which is 7.1% for data science developers slightly higher than the 6.9% for other developers — a 2.7 % difference.

Based on the above I wanted to find out whether the developer type (data science or other) can predict job satisfaction in a statistically significant way. For this reason I conducted a Chi-Square Test for Independence ,which confirmed with 95% confidence, our notion that data science developers have a statistically significant higher, on average, job satisfaction.


Conclusion

In this article, we examined some of the differences between developers active in the data science field compared to other developers using the Stack Overflow 2017 survey data.

  1. We calculated that they represent 12% of the total, including developers that may be active in other fields but listed a data science related job description.
  2. We took a look at their working habits as represented by working remotely and programming outside of their work and found that data science developers display higher percentages of remote working and programming activity outside of work including contributing to open source.
  3. Finally, we looked at the job satisfaction ratings between the two categories and concluded that there is a small (2.7%) on average but significant and persistent difference in favor of the data science developers. Performing a statistical test we confirmed that the developer type can predict the job satisfaction with a 95% confidence level (alpha=0.05).

The findings are interesting and have been initially confirmed by statistical testing but are still mostly observational, and a more rigorous study should be performed to reach final conclusions.

The full analysis including the code can be found at the GitHub repository available here.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade