What we’re looking for (and what we’re not looking for)
We’re really interested in the process you take to tackle this task and how well you can communicate your findings. We’re not interested in the findings themselves so please don’t worry if you’re not able to draw any concrete conclusions at the end of it. A finding that you need more data or that more investigation is required is perfectly valid. If you can explain what you’ve done and why you are or are not able to draw conclusions from your work that’s the important thing.
Data Analyst Assignment – UK

Completing the task
When you’ve finished the task please upload any files that are needed to understand your final results or how you got there with the link you have been provided. We definitely want to see your working (e.g. SQL queries and data manipulation) as much as your final insights so if they’re in separate files send them all through. A zipped folder is ideal if possible.

We know everybody has different commitments and unforeseen things come up all the time so if you are unable to finish this at the expected time just let us know and we can make arrangements for you to submit at a later date.

Getting help
If you get stuck at any point whether in connecting to the database under standing what the data represents or interpreting our expectation please feel free to contact owen@dice.fm and we can answer any questions or jump on a call to troubleshoot or explain further.

The task
What we would like from you is to understand more about retention of DICE fans so we would like you to use the data provided (explained below) to answer the question “are some segments of DICE fans retained better than others?”

In answering that question, we’d like you to address these points:
● How did you define retention, and why did you make that choice?
● If you see differences in retention between groups what do you think might explain that? If you do not see any differences why do you think that is? You don’t need to prove any of these ideas with data.
● If you had more time and more data what other segments would you like to look into. What other data points might be required to do this?
You can take any approach you like to tackle the task and can use whatever technology suits you best to present your results (options include slide presentations R Mark down note books Jupyter notebooks or spreadsheets). Here are a few pointers though:
● We’re interested in your ability to work with SQL so even if it’s possible to extract all data from the database and manipulate it elsewhere we’d advise against doing so.
● Simple segmentation is enough – for example age groups cities or fans who did/did not perform a given action. There’s no need to train an ML model to segment users as you’ll only be working with limited data and we’re more interested in your ability to communicate insights than to use complex models.
● There’s no need to exhaustively test every possible segment. We’ve given you a few fields so that you can select what you think is likely to be most important. Don’t worry if the ones you investigate first don’t reveal any clear trends – we’ll be more interested in why you chose the approach you did, rather than the outcome.
● We’d love to see a few visualisations of the data to help get your points across so please do include a few charts that you think communicate the most important points effectively.