During my spring break Princeternship, I intended to learn more about careers in statistical analysis by observing an alumnus at Rentrak Corp. Rentrak is a leading media measurement company, and the top competitor to Nielsen. The main product is a constantly updating database of tracked viewing numbers of TV shows and movies across different demographics and time slots. I expected to shadow a lot of roles concerning research and data gathering. What I found out was the job was actually a highly integrated combination of primarily software development, programming, and research.
On my first day at the job, I was introduced to Princeton alum Dwayne Thomas ’04, who was a Quality Assurance Statistical Analyst. He, and several other people, had jobs testing the database that the programmers produced. Testing “issues,” or specific new additions to the program, the Quality Assurance team members used Kanban, an efficient workflow tool, to implement agile software development, which focuses on iterative and incremental changes to the program. Kanban is a visual process that limits the number of items in a person’s workflow at a single time-step, which Rentrak employees visualized using a third-party project-tracking software called JIRA. JIRA itself was a database holding records of all the tests they performed on the Rentrak database.
I was also introduced to several Configuration Management team members, who had the task of executing computerized tests of the Rentrak database. Whereas Quality Assurance provided real-person analysis of add-ons to the software, Configuration Management consisted of programmers who ran automatic tests, such as whether the database was updating at a regular pace, or whether each link on the database page correctly linked to another one.
Next, I observed the programmers in the Software Development division. The database managers pulled large amounts of raw data from broadcast companies and on-demand services, and to simplify it and fill in gaps natural to raw data. The backend programmers used C++ to integrate the data into Rentrak’s database, working individually, but combining their work to produce “millions of lines of code.” The front-end programmers used Perl to create the look of the company website and of the database system.
I also met the researchers, who acted as the step in between database and backend programming. The engineering managers would feed the researchers simplified, but un-analyzed data, and the researchers would use statistical programs such as SAS and R, to adjust the data for certain factors, such as disproportionate demographics sampled or time-zone adjusts to ensure statistical accuracy.
This Princeternship experience not only exposed me to careers in statistical research, but also to a career as a software developer or software tester. This Princeternship was valuable in that it showed me that tasks in the real world are not so clear cut – developers and researchers could spend most of the day working together, and yet, still have separate but integral roles.