登录
  • #面试经验
  • #数科面经
  • #分析|数据科学类
  • #facebook

FB Data Science - Analytics挂经(全新题)

yhy7163
9118
31
刚面的Facebook DS岗Product组,碰到鬼了,面试官全程黑脸,我答的不满意他就说我要的不是这个。

2020-1月末,Video Interview。

SQL:

### Table 1: searches

### columns:

* date (string), date of the search

* search_id (int), the unique identifier of each search

* user_id (int), the unique identifier of the searcher

* age_group (string), ('U30', '30-50', '50+')

* search_query (string), the text of the search query

Sample Rows:

date | search_id | user_id | age_group | search_query

-----------------------------------------------------------------------------------------

‘2020-01-01’ | 101 | 9991 | 'U30' | ‘michael jackson’

‘2020-01-01’ | 102 | 9991

本帖隐藏的内容需要积分高于 188 才可浏览,点击前往一亩三分地论坛阅读

hat users need to find.

1. How to implement such a function? What to look at? What features or data would you use to try to predict what the users are typing?

2. How to evaluate this? How do you know if our suggested result is good or not?

3. Imaging that we are trying to insert external website link to the searched results (wikipedia, IMDb, etc.), how do you measure if this is a good idea?

我从地里的帖子自己总结了一个模版,但是似乎面试官并不买账,疯狂follow-up,我跟他clarify,他重复了一遍问题然后说让我想,而且我没get到他到底想问什么,问他是不是可以用A/B testing来判断是不是good idea?他说你用什么metrics?我说DAU, Conversion rate之类的,看他也没什么反应。

全程几乎没有什么提示,他有时候记一下笔记,估计我每答错一个他就记下来,anyway,大型翻车现场。

求大米。

谢谢,祝好运。
31条回复
热度排序

发表回复