The take home assignment: Analyze data set X and create a presentation of your findings
This is a common task that seems easy on teh face of it but is actually very difficulyt. Ther are two main dangers: 1. Since there are very few parameters you may go in a direction that is not what the interviewers are looking for. 2. If you are given an ambiguous task you may want to clarify things. This can either be interpretted as good (showing initiative) or bad (not being independent). If you do get a task like this, have someone else read it if possible and interpret it without you telling them what you think it is about. That way you will at least bet some feedback on what a common interpretation is. These types of assignments are good for graduate students as they are used to the format from doing homework and generally have more time to spend on them than people who are working already.
Questions that are easy for experienced people but difficult for people who just put "R programming" on their resume becasue it sounds good
- Fizzbuzz type questions
- What is the code to convert long format to wide format data in R/Python?
- What is the code in R/Python to find the mean score for several groups?
- Qhat is the SQL code to query a table for US States starting with "N"?
Questions that aim to figure out whether you "fit" with an organization
- Name a data scientist whose work you admire, why do you admire this work?
- How do you know if an idea you have is a good one?
- In your opinion, how does data science differ from data analytics and machine learning?
- What compelling examples do you know of for how data has helped education?