每日一练 | Data Scientist & Business Analyst & Leetcode 面试题 883

点击上方 蓝字 会变美

“

Data Application Lab 自2017年6月15日起，每天和你分享讨论一道数据科学(DS)和商业分析(BA) 领域常见的面试问题。

自2017年10月4日起，每天再为大家分享一道Leetcode 算法题。

“

DS Interview Question

Why data cleaning plays a vital role in analysis?

BA Interview Question

Write a query in SQL to obtain the name and position of all physicians who completed a medical procedure with certification after the date of expiration of their certificate.

LeetCode Question

Best Time to Buy and Sell Stock

Description:

Say you have an array for which the ith element is the price of a given stock on day i.

If you were only permitted to complete at most one transaction (ie, buy one and sell one share of the stock), design an algorithm to find the maximum profit.

Input: [7, 1, 5, 3, 6, 4]

Output: 5

DS Interview Question & Answer

Design an experiment to figure out which web design alternative to use. Assume there have been no other experiments done and there is no knowledge of the user behavior. Discuss potential issues that can occur with the conclusions and how to avoid them.

Identify the nature of the change and the metric to consider to decide which version of the site to choose

Next, decide the number of samples/visits necessary to hit the necessary statistical significance (e.g. 95%). This can be done by using a chi­-squared test (if we are using a binomial random variable of clicking vs. not clicking) or a z­test (if we are using a normally distributed random variable). You can then evaluate the p­-value to identify whether the metric of the B test is statistically significantly different than the metric of the baseline A test. If it is and the metric is better than the baseline, then the alternative site is the better way to go.

Some other issues you should consider in this answer:

1) Identify potential biases due to interactions across pages. Talk to the product manager and see if there are ways that a random sampling may not work to test the nature of the change you’re proposing for a web page.

2) Perform a A/A test which implies testing two random samples of visitors, and check if the distribution and metric of choice does not have a statistically significant difference. This will ensure the fairness of the A/B test. An A/A test ensures that your audience doesn’t have a particular skew or bias and a randomized selection for an A/B test will be statistically relevant

BA Interview Question & Answer

Write a query in SQL to obtain the names of all the physicians, their procedure, date when the procedure was carried out and name of the patient on which procedure have been carried out but those physicians are not certified for that procedure

Answer:

SELECT p.name AS "Physician",
pr.name AS "Procedure",
u.date,
pt.name AS "Patient"
FROM physician p,
undergoes u,
patient pt,
PROCEDURE pr
WHERE u.patient = pt.SSN
AND u.procedure = pr.Code
AND u.physician = p.EmployeeID
AND NOT EXISTS
( SELECT *
FROM trained_in t
WHERE t.treatment = u.procedure
AND t.physician = u.physician );

Sample Output:

Physician              |       Procedure                        |        date                     |  Patient
--------------------------+---------------------------------------+-----------------------------+----------------------------
Christopher Turk    | Complete Walletectomy         | 2008-05-13 00:00:00  | Dennis Doe
(1 row)