Andreas Weigend | Social Data Revolution | Fall 2014
School of Information | University of California at Berkeley | INFO 290A-03

Table of Contents

Homework 1: LinkedIn

Assigned: Tue Sep 30, 2014
Due date: Mon Oct 6, 2014 at noon
Maximum number of points: 10 (out of 100)
URL of form for submitting your homework:

Like many social networks, LinkedIn collects a lot of data about the user, refines these data into meaningful metrics, and creates valuable services and products as an end result. In this simple exercise, we ask you to put yourself in LinkedIn’s shoes and think about the business logic of LinkedIn and the way in which it utilizes data.

If you do not yet have a LinkedIn account, we suggest that you spend some time to create one and poke around the user profile settings to get familiar with things. You do not have to use your real personal information if you feel uncomfortable. However, you will get much better result if LinkedIn provides real recommendation for you.

Open up your own LinkedIn profile and consider the questions below. Try to explain your answers in greater details and be clear about assumptions and prior knowledge that you bring to the response. If you feel that you do not have the necessary knowledge or data to respond to the question, describe what you think these are and how one might acquire the necessary data. Your answers will be evaluated not just based on what you already know, but the way in which you approach problems that you do not already know the answer to. Try to think through the problems logically and give clear concise answers.

  1. Of all the different fields that LinkedIn suggests you fill up, what is the most surprising? Why do you think LinkedIn is interested in knowing the answer to that field?
  2. Look at the “People Also Viewed” column to the right. Can you think of why people would view these profiles together with yours? What can LinkedIn learn about you from the fact that these people are somehow associated with you? (E.g. did you attend a certain conference? Was in a certain school club?)
  3. Scroll through the “People You May Know” section.
    1. Is the recommendation system doing a good job? What makes you feel that way?
    2. What does LinkedIn know about you that allows it to recommend these profiles to you? What do you think are some of the attributes used by LinkedIn to rank and prioritize recommendations?
    3. How would you design a recommendation system for LinkedIn profiles? What are the key variables that you think are essential? How would you measure/gather these data?
    4. What do you think are the tradeoffs that LinkedIn makes in determining which profiles to recommend to the user?
  4. One of the toughest problems for companies like LinkedIn is reputation. Recruiters are constantly on the lookout for the best potential employees and LinkedIn tries to provide features that help recruiters do that. We look at two different parts of the profile that attempt to fulfill this purpose and consider their effectiveness.
    1. On the right side, there is a “Profile Strength” ranking that looks like this:
      How do you think LinkedIn is evaluating a user profile? (E.g. do you think profiles are rated on an absolute scale based on meeting criteria or on a normalized scale relative to other user profiles?) What are the metrics and criteria that you would use to rank a user profile and why do you think that they are important?
    2. On your main user profile page, there is a “Skills” section that looks like this:
      These are endorsements for a user’s listed skillsets by people who are connected to the user. Do you think that this is an effective way of evaluating an individual’s skills? What are the benefits and drawbacks of this peer endorsement system? Do you have ideas to improve this system or to create a better way of evaluating a user’s skill set?
  5. Going forward, we can also think of LinkedIn’s collection of user data as more than just a recruitment or professional networking database. One of the solutions that LinkedIn provides is the Linked Sales Navigator, which helps sales people find business opportunities through the LinkedIn social graph.
    1. Given what you know from your interactions with your LinkedIn profile and the data that you think LinkedIn has access to, what are some other possible products or services that you think LinkedIn can offer? (E.g. a professional dating service?) Justify your answer by explaining how specific types of data that LinkedIn has access to can provide a competitive advantage in the service that you plan to provide.
    2. Are there data that LinkedIn does not presently have access to that can greatly improve either its existing services, or the hypothetical services that you suggested? How should LinkedIn go about obtaining these data? Is there value that LinkedIn can provide to the end user in exchange for these data?

Direct any questions to Raven Jiang (raven at cs dot stanford dot edu)