Do engaged new coders pay more for courses?

I recently participated in a Data Science Bootcamp course organized by IBM’s Big Data University and Ryerson University in Toronto taught by Prof. Murtaza Haider. Great experience and amazing inspiration! I decided to put my new data science skills into action to look deeper into behavior of new coders as I am a new coder exciting to learn more.

I decided to answer a question: Do new coders who spend more time learning to code pay more for coding courses?

There are so many free courses that will help you build your coding skills. Can companies like Coursera or edX increase their revenues by encouraging students to invest more time in learning how to code? What are the other factors that will make it more likely for a new coder to pay for coding classes?

The data that I used to answer this question comes from the 2016 New Coder Survey conducted by freeCodeCamp among 15,000+ people who are new to software development. I used R for basic analysis.

What did I learn about an average respondent (based on mean analysis)?

He:
Is 31 years old
Is employed
Is male
Has debt
Spends 13 hours each week on learning how to code
Has earned $45k last year
Has been programming for 29 months so far (4.4 years)
Spent $1,154 on learning how to code so far

Most of the respondents spent $0 or very little on learning how to code (see Figure 1). This is in line with popularity and availability of free courses.

Figure 1: The amount of money spent on learning how to code so far.

Based on the initial exploratory analysis I decided to limit the number of the analyzed factors that are likely to influence the amount of money spent on coding to four:

the amount of time spent each week on learning how to code, programming experience in months, income from previous year, and age (see Figure 2 for plotting of these factors against the amount of money spent so far on learning how to code).

Figure 2: Plots for the four factors potentially influencing the amount of money spent on learning how to code.

I applied regression analysis to see how the different variables contribute to the amount of money that the new coders spend on learning (see Figure 3).

Figure 3: Regression analysis. Comparison of four models.

I decided to select model 3 to interpret my results. This decision has been influenced by the analysis of the scatter plots which shows that the relationship between age and money spent on learning how to code would need to be analyzed in several age brackets (I’ll leave it for step 2 in my analysis).

What did I learn from this analysis?

  1. All else being equal, with each additional hour spent on learning how to code per week a new coder spends $29 more (in total) on learning to code.
  2. All else being equal, with each additional month of programming experience a new coder spends $8 more (in total) on learning to code.
  3. All else being equal, with each additional $1,000 of income earned per year a new coder spends $8 more (in total) on learning to code.

What is the answer to my main research question?

YES, new coders who spend more time learning how to code spend more money on coding courses.

What are the implications for companies that provide coding courses?

To increase revenues, look for ways to encourage new coders to spend more time on learning.

Thanks for reading! Did you find this story useful, inspiring? Like, share and comment.

About the photo: I took it at the Computer History Museum in Mountain View in May this year. Strongly recommended for anyone interested in computer science!