Structure Approach to Analytics Interviews
•Download as PPTX, PDF•
2 likes•685 views
This is the talk I gave to Udacity Nanodegree students
Recommended
Recommended
More Related Content
What's hot
What's hot (11)
Viewers also liked
Viewers also liked (20)
Similar to Structure Approach to Analytics Interviews
Similar to Structure Approach to Analytics Interviews (20)
More from Ankit Jain
Recently uploaded
Recently uploaded (20)
Structure Approach to Analytics Interviews
- 1. A Structured Approach to Analytics Interviews ANKIT JAIN
- 2. Who am I? • Data Scientist@Uber • Previously, led Data Science@startup in India • Worked at Facebook, BofA • UCB, IIT Bombay (India) • Half Marathon Runner • Husband
- 3. Why to enter into Data Analytics/Science? 2.5 *1018 Bytes of data generated everyday (--IBM) 150,000 Shortage of Data Scientists in US alone in 2018 ( -- Mckinsey and Co.)
- 4. Analytics Interview is not Rocket Science
- 5. What are Employer’s Requirements What are Employer’s Requirements? • Technical Skills • Knowledge of Statistics • Structured Thinking • Business Understanding • Communication Skills • Attention to details
- 6. Types of Interviews Analytics Interview Technical Behavioral Case Interviews
- 7. Technical Interviews • SQL : Joins, GroupBy , Where, Having, Window Functions, Query optimizations, NoSQL • R/Excel : Level depends on the job • Tableau or other BI tool: Useful if you are going for reporting/dashboarding kind of role
- 8. Guesstimates (Type 1 Case Interviews) • These are the kinds of interviews where you are required to estimate a number by putting a framework to the problem, making assumptions and coming up with a final number. • For e.g. “ Estimate the number of balls/ liquor bottles in USA” • Few pointers: • It is the approach that matters – not the final numbers • Call out all the assumptions and possible blind spots • Clearly state the framework and segmentations you make
- 9. Real Life Case Study A case interview starts with a broad business question and narrows down in a particular direction. It generally starts with a very simple problem and then interviewers tend to increase the complexity as you solve the problems. Few pointers: • Case study is all about three things : Structure, Structure, Structure. Focus on putting a frame work to the problem. • Again call out all the assumptions. • Think out loud, interviewer is looking for your thought process • Communicate effectively
- 10. Some more things 1. Update your Linkedin profile 2. Attend local meetups 3. Keep yourself updated through: 1. News sites like www.datatau.com (hacker news for data science/analytics) 2. Follow data science/analytics leaders on Twitter ( for e.g DJ Patil, Andrew Ng etc. )
- 11. CASE STUDY EXAMPLE Source : https://modeanalytics.com/
- 12. Yammer ($1.2Bn Acquisition by Microsoft) Yammer is a social network for communicating with coworkers. Individuals share documents, updates, and ideas by posting them in groups.Think of it like Facebook for Work. Yammer has web,mobile and tablet apps.
- 13. Investigating a Drop in User Engagement What caused this dip in engaged users?
- 14. Prior Hypothesis • Holiday • Broken Feature/App • Error in Data Recording Script • One Time Marketing Event • Etc…
- 15. User Growth on Yammer Platform Growth Rates look normal. High on weekdays and low on weekends
- 16. Engagement of Old Users
- 17. Engagement By Device Major Engagement Drop for Mobile Users
- 18. Zooming in on Mobile Users
- 19. Further Exploration It seems the issue is with the mobile app related to long-time user retention. Possible issues can be: 1. Mobile app has become too heavy with new features and it crashes 2. Re-engagement Emails that were sent for retaining users are not been viewed on mobile clearly 3. A bug in recording data in mobile 4. Etc …
- 20. Thank You
Editor's Notes
- So it’s likely that there’s a problem with the mobile app related to long-time user retention. At this point, you’re in a good position to ask around and see if anything changed recently with the mobile app to try to figure out the problem. You might also think about what causes people to engage with the product.
- 1. One of the easiest things to check is growth, both because it’s easy to measure and because most companies (Yammer included) track this closely already. In this case, you have to make it yourself, though. You’ll notice that nothing has really changed about the growth rate—it continues to be high during the week, low on weekends:
- Since growth is normal, it’s possible that the dip in engagement is coming from existing users as opposed to new ones. One of the most effective ways to look at this is to cohort users based on when they signed up for the product. This chart shows a decrease in engagement among users who signed up more than 10 weeks prior:
- The understanding that the problem is localized to older users leads us to believe that the problem probably isn’t related to a one-time spike from marketing traffic or something that is affecting new traffic to the site like being blocked or changing rank on search engines. Now let’s take a look at various device types to see if the problem is localized to any particular product:
- So it’s likely that there’s a problem with the mobile app related to long-time user retention. At this point, you’re in a good position to ask around and see if anything changed recently with the mobile app to try to figure out the problem. You might also think about what causes people to engage with the product.