Slide from a presentation I did back in late 2015/ early 2016. Our (Wego Data Science team) primary data analysis setup then was relying on Google Sheets and App Scripts to process and deliver data stored on BigQuery. We’ve come a long way since then, and it is no longer our primary means of scheduling processing and delivering data, but the pointers are still very relevant.
It is a good setup for small and new data teams who want something quick and easy to deploy (much lesser coding required). In the long run I’d still recommend a proper setup though. At the end of the day, it is a free tool with many limitations which would not go well with growing data demands.
Pardon the aesthetics, I stripped the formatting out from the original presentation to avoid organisational sensitivities.
One thing I did not mention in the slides of the benefits of using Google Sheets as a transport/ holder of data is that it provides a very visual interface for account/ business managers to do troubleshooting on their own, as they might not be technically inclined/ have the time to go through lines of code if information produced does not look right. More on that in future.
Google Sheets + App Scripts
Pros:
- Visual interface for checking steps and taking in user input
- Fast to deploy and schedule with no overheads or systems to maintain
- Lesser coding required (might not be a valid excuse, more of a quick fix)
- Can create own script libraries to use across different projects
Cons:
- 6 min limit per run (although you can script a workaround)
- Reliability sometimes an issue
- 2 million cells limit. Although I don’t think you should be printing out the output if your data is in that range
- Schedules created stay at the user level (rather than sheet level). Can be a pain in checking what is being scheduled by teammates
- Less control and flexibility in all aspects compared to creating your own data processing pipeline (with Python and Compute Engine, for example)
Is the above information useful? Would you like more details on a specific aspect? Let me know in the comments below or drop me a message at aaron@aaronteoh.com. Thanks for reading!