Question

3) a) Describe different routes to aggregation using as clustering principles user attributes (e.g. age, gender, location) and activity type (liking, following, tagging). What does aggregation of such data achieve? How are the aggregated data deployed to promote the business objectives of platforms? b) Use the case of Facebook likes to illustrate different routes to aggregation and how the data that result from this aggregation are used to promote Facebook’s business objectives

Answer 1

solution:

given data:

aggregation using as clustering principles user attributes:

3)

a)

• Social data is the computable data footprint of user participation in social media.
• It is produced through the far-reaching standardization of social interaction that allows users to perform daily and en masse hyper-stylized activities such as ‘tagging’, ‘following’ or ‘liking’.
• The data thus produced are piled up and processed. The outputs of these operations (e.g. similarity, popularity or trending scores) are carried back to users in the form of personalized suggestions, embedded into platform functioning and shared with third parties through APIs and other boundary spanning technologies, reinforcing the process through which social interaction is made the engine of social data production.
• These conditions set social data apart from data generated through automated technologies of data tracking and recording, and outline the distinct nature of the contribution it makes to the developments associated with big data.
• Clustering attempts to segment items so that members of one group are more similar to each other than to members of other groups.
• These items are most often customers, but can also be products, patients, prescriptions, phone calls, emails, or any other item relevant to the enterprise.
• Clustering algorithms do the segmentation by analyzing the characteristics of the items and finding the best ways to group them by similarities.
• Clustering is an important step in the process of data analysis with applications to numerous fields. Informally, clustering is defined as the problem of partitioning data objects into groups (clusters), such that objects in the same group are similar, while objects in different groups are dissimilar.
• Clustering then becomes the problem of grouping together data objects so that the quality measure is optimized.
• clustering that is based on the concept of aggregation. We assume that given a set of data objects we can obtain some information on how these objects should be clustered. This information comes in the form of m clusterings C1, . . . , Cm. The objective is to produce a single clustering C that agrees as much as possible with the m clusterings.
• Clustering aggregation as the optimization problem where, given a set of m clusterings, we want to find the clustering that minimizes the total number of disagreements with the m clusterings.
• Clustering aggregation provides a general framework for dealing with a variety of problems related to clustering: (i) it gives a natural clustering algorithm for categorical data, which allows for a simple treatment of missing values, (ii) it handles heterogeneous data, where tuples are defined over incomparable attributes, (iii) it determines the appropriate number of clusters and it detects outliers, (iv) it provides a method for improving the clustering robustness, by combining the results of many clustering algorithms, (v) it allows for clustering of data that is vertically partitioned in order to preserve privacy.

b)

• Customer relationship management (CRM) has become widely recognized as an important business approach defines CRM as an “enterprise approach to understanding and influencing customer behavior through meaningful communications in order to improve customer acquisition, customer retention, customer loyalty, and customer profitability”.Hence,customeracquisition(or:identification) can be seen as the first step in a customer relationship management cycle that, together with retention and customer development form a complete cycle geared at creating a better understanding of (potential) customers in order to increase long term customer value to the firm.

• Online profiling can be divided into two categories: reactive and non-reactive data collection.Non-reactive data collection focuses on the collection of data concerning Web usage behavior, e.g., IP addresses of visitors, timespent on certain Web pages, and clicking behavior information. These data are used to gain insight into Web user behavior, and thus, characteristics of individual visitors or visitor groups. Non-reactive data form a large and potentially interesting source of online profile information. However, for the construction of online user profiles, the observed usage behavior must first be trans-formed into meaningful variables. The construction and definition of such variables is not always a easy task. In our study, we therefore primarily focus on the retrieval and analysis of online profiles based on reactive rather than non-reactive data.

• Reactive data collection zooms in on visitor characteristics which cannot be collected through tracking Web usage behavior of a visitor. Instead, reactive information is collected by using forms and selection menus, which have to be filled in by visitors themselves. Reactive data requires little to no recoding of the original variables and they are immediately collected at the user level. Moreover, in the case of Facebook, providing reactive data requires very little effort from the users. For example, when joining Facebook, users are asked to provide certain personal background information (e.g., name, gender, date of birth).Users provide this information by selecting the appropriate options.This basic background information can be supplemented by more personal information concerning, for example, hobbies, relationship status etc. Finally, by “liking”other pages, personal preferences for persons or objects can be indicated.

• The resulting online profiles can be of great value for marketeers, as they can be used to identify different (segments of) users (customers) that require different marketing approaches. Moreover, it enablesthe company to know its potential customers, that is, individuals that indicated a preference towards the product/brand by “liking”it on Facebook.

• Facebook users put personal information on their Facebook page.Some examples are someone's name, gender, date of birth, e-mail address, sexual orientation, marital status, interests, hobbies, favorite sports team(s), favorite athlete(s), or favorite music. Furthermore, it is possible to specify your Facebook friends, post messages, publish pictures or other content. Consequently, the potential value to marketeers and researchers of the information available through Facebook is substantial.

• The Facebook data collection framework that we propose consists of three steps: 1) identification of “fans”of the Facebook page, 2) retrieval of relevant data for the identified fans, and 3) preparation of the data.The first step of this framework requires administrator rights to the page, in the other steps public information from the relevant pages needs to be collected

• he owner of a Facebook page is in principle the administrator of the page. Personal pages are typically managed only by the page owner, however, in the case of a company's Facebook pages, the page administrator can also give other Facebook users these administrator rights. It is possible to have multiple administrators for one Facebook page, e.g., multiple marketing and CRM employees may be page administrators. As Facebook page administrator, one has certain privileges in comparison to regular users or visitors of a Facebook page. As administrator, one can edit, publish and withdraw content, target advertisements and install Facebook apps on the page. Also, administrators have access to Facebook Insights, a dashboard which provides statistics on user's growth, demographics, consumption of content, and creation of content. However, the information made available through the dashboard is aggregated over users who “liked”the page. Consequently, the possibilities concerning the analysis of individual specific data using this feature, are limited.

• On Facebook, users can indicate whether they “like”another Facebook page.Thus,they are able to express a form of affinity with the company.

• To gather a Facebook user's profile information relevant to customer relationship management and/or for marketing purposes, the information resources must be combined. For convenience we introduce the term “fan”for users who “liked”a Facebook page. In fact, Facebook itself originally gave users the option to “become a fan of” other pages and changed this into “like”. The data collection framework consists of three parts: 1) identifying the ‘fans’,2) gathering the personal information, and 3) preparing and structuring the gathered data

• A user's location is represented by a string with the name of the city or town someone lives in and/or comes from, together with a URL to the Facebook page of that location. Location may be useful when analyzing fans of a page and we may be interested in more details about the location. In particular, for a geographical overview of the fanbase one needs to know the country, continent, and the coordinates (latitude, longitude) of a location. This can be achieved by using the GeoNames geo graphical database . The GeoNames API has a fuzzy search engine which accepts all kinds of input.

• A customer relationship manager or a marketing manager would like to make these data operational by,for example, investigating whether fans can be segmented according to their indicated preferences and/or background characteristics. That is, is it possible to identify groups of fans on the basis of individual specific like data.

• There exist several methods for clustering high-dimensional data.One popular approach is to use a two-step procedure. In the first step, a dimension reduction technique is used to reduce the dimensionality of the data. In the second step, cluster analysis is applied to the data in the reduced space.

• K-means clustering,finds clusters by minimizing the sum of squared deviations between the individual observations and their cluster means.

• Apart from Google, Facebook is probably the only company that possesses this high level of detailed customer information. The more users who use Facebook, the more information they amass. Heavily investing in its ability to collect, store, and analyze data, Facebook does not stop there. Apart from analyzing user data, Facebook has other ways of determining user behavior.

• Tracking cookies: Facebook tracks its users across the web by using tracking cookies. If a user is logged into Facebook and simultaneously browses other websites, Facebook can track the sites they are visiting.
• Facial recognition: One of Facebook’s latest investments has been in facial recognition and image processing capabilities. Facebook can track its users across the internet and other Facebook profiles with image data provided through user sharing.
• Tag suggestions: Facebook suggests who to tag in user photos through image processing and facial recognition.
• Analyzing the Likes: A recent study conducted showed that it is viable to predict data accurately on a range of personal attributes that are highly sensitive just by analyzing a user’s Facebook Likes. Work conducted by researchers at Cambridge University and Microsoft Research shows how the patterns of Facebook Likes can very accurately predict your sexual orientation, satisfaction with life, intelligence, emotional stability, religion, alcohol use and drug use, relationship status, age, gender, race, and political views—among many others.
• The simplest explanation for this is that Facebook uses that data to make money. No, Facebook doesn’t sell your data. But it does sell access to you, or more specifically, access to your News Feed, and uses that data to show you specific ads it thinks you’re likely to enjoy or click on.
• This targeted advertising is big business for Facebook. The company reported advertising revenue of $40 billion last year, and it’s only going to keep growing. Given the company’s recent privacy issues involving Cambridge Analytica, a third-party data firm that got its hands on personal data for as many as 87 million Facebook users without their permission, we thought it might be helpful to take a closer look at how Facebook uses your data to make money.
• Outside businesses can collect your data if you grant them permission — for example, if you use your Facebook account to log in to a third-party app like Uber or Spotify. Facebook just announced changes to some of those data-sharing APIs to better ensure that agreeing to share your own data won’t let those outside companies collect data about your friends without their permission. (This is what happened with the Cabridge Analytica situation.) Of course, any data you share publicly to your Facebook page is accessible to anyone online.

please give me thumb up