60 minute live coding. There were two CSVs, collaboration and user, and I could pick my own environment to do the analysis in pandas.
Landing source distribution
Analyze which entry points non-creator users came into the page from. The interviewer followed up on whether it should be computed at the event level or the user level, and on whether highly active users generating lots of repeated events would bias the result.
Most frequent collaborator
Find the person each user collaborates with most often. First you have to define collaboration: two people who edited the same page. Then you generate user pairs, deduplicate them, and decide whether to measure collaboration strength by the number of shared edits or by the number of distinct shared pages.
Cross-country collaboration
Compare how likely users in different countries are to collaborate. The point is not to just give one number, but to be explicit about the numerator and denominator, for example the share of users who collaborate, shared pages per user, or a collaborative editing rate.
Discussion
Loading comments…