User Guide
Thank you very much for participating in this study. Big Sister is a prototype of a smartphone app that makes profiles and song recommendations based on texts gathered from your social media accounts or by processing a text that you enter yourself.
This brief document will guide you on how to install and use the Big Sister app, explaining its main functions and containing instructions on troubleshooting any problems you might have. This guide also includes a glossary of terms in case you are interested in learning more about how Algorithm Procedure or formula for solving a problem or accomplishing some end, based on conducting a sequence of specified actions. In computer systems, an algorithm is an instance of logic written in software to effectively produce or compute an output from a given input [See Input/output].ic Profiling It is the use of personal data (your clicks, tweets, likes, or posts, amongst many other data) to evaluate certain aspects related to your persona and to create a profile that informs recommendations and other personalised content. The purpose of profiling is to predict the individual’s behaviour and take decisions based on these predictions. You might have experienced profiling on Facebook, for example: it informs the content you see in your news feed. and recommendations are generated by Big Sister and the most popular digital platforms.
1. Download and install Big Sister
You should have received an email inviting you to test Big Sister from Google or Apple depending on which platform you previously choose to test the app. Click on the link and you will be directed to the PlayStore or AppStore.
1.1 Android
Download the app and wait for it to install on your phone. You will need a Gmail account to install the app, as Google requires users to have a Gmail account to test apps on the Play Store.
1.2 iOS
To install Big Sister, you will need to install Apple’s TestFlight app on to your iOS device. You can find TestFlight app in the AppStore or click here. You will then be able to download the Big Sister app and install it on your device.
2. Gathering information about you
Click on the Big Sister Icon on your phone. You can change the language (English or Spanish) by clicking on the globe icon. After the introductory message, Big Sister will ask you to click on the eye icon to start a new test. Then, Big Sister will ask you to provide an input [See Glossary] or source of information that it will use to analyse and describe your identity. The Big Sister app works with three sources of information: your Facebook account, your Twitter account, or open texts about a public issue or figure. It only needs one of these sources to work, though you might want to try all three!
2.1. Facebook
Since 2018, when it became known that a firm called Cambridge Analytica had been misusing Facebook data to target political advertisements, Facebook has made it harder for apps to access your personal data. If you’d like to use Facebook as a source for Big Sister, we ask you to send us your Facebook username so we can manually add you to their app testing program. Go to your Facebook profile and copy the link, your username is what continues after the www.facebook.com/. Then you must check the notifications on your Facebook account, preferably in a desktop, and give the authorization to the Big Sister app.
Note: We do not store your username or password and have no access to your account. This information will only be used to automatically confirm your authorisation.
2.2. Twitter
To provide access to your Twitter data, Big Sister will direct you to your web browser, where you will be asked to log in to your Twitter account. This log in authorisation is provided by Twitter, not by Big Sister.
Note: We do not save your username or password and have no access to your account. This information will only be used to automatically confirm your authorisation.
2.3. Open text
If you do not have a Facebook or Twitter account, or you do not wish to provide access to their Facebook or Twitter data, Big Sister allows you to provide another kind of input: a text written by you. When you select this option, Big Sister will prompt you to write something about a public figure or a particular issue. It will then use that text as the basis for analysis and recommendation.
If you would like to receive a personalised result, we recommend that you use a text that you have written by yourself. But we also encourage you to play with this feature by inputting texts written by others. Perhaps you’ve always wondered just how ‘impulsive’ a dictionary is, or whether Miley Cyrus is actually ‘conservative and traditional’. Now you can find out!
Tip: The minimum number of words required for analysis is 100. According to the developers of IBM Personality Insight, more words are likely to produce better results.
2.4 Logout Icon
By clicking in this icon you can see Facebook, Twitter or Spotify accounts that you have given your permission to access your data. There you can choose one to log it out, cancelling the permission to access your information for future tests.
Tip: Once you give the app permission to collect data from your social networks (Facebook or Twitter), it will not be necessary to replicate the same process in future tests.
3. Processing your results
Big sister uses different Application programming interfaces (API) A set of clearly defined definitions, protocols and methods to establish a communication between different systems or components of a system. For example, our Big Sister app automatically makes a request of information (tweets) to Twitter’s database through Twitter’s API. If this request is valid and has been authorised by the user, the API sends data in response, like a user’s tweets. [See Glossary] to gather and process information. First, depending on the selected source of data, the app sends a request to the API of Facebook or Twitter to receive your latest posts or tweets in a textual and anonymized format. Then, Big Sister app sends the gathered data as inputs [See Glossary] to Watson Personality Insights API. This service uses an Algorithm Procedure or formula for solving a problem or accomplishing some end, based on conducting a sequence of specified actions. In computer systems, an algorithm is an instance of logic written in software to effectively produce or compute an output from a given input [See Input/output]. developed using a theory, drawn from psychology, marketing, and other fields, that human language “reflects” or “provide windows” into the personality. Specifically, this service seeks to infer the users’ personality through the analysis of how they write in their social media accounts with Linguistic analytics A method of computational science that seeks to understand human language as it is spoken or written using algorithmic models. Linguistic analytics uses data a person generates through blogs, tweets, forum posts and email to generate a score relative to a sample population along a spectrum of cognitive and social characteristics. [See Glossary]. It calculates the frequency and usage of certain words and compares them with scores from surveys that were conducted among thousands of individuals along with data from their Twitter feeds. When data processing is complete and test results are received, a button will appear to see your results.
4. Visualizing and comparing your results
This is this screen you can see what Big Sister think of you. Three main types of results are shown:
4.1. Personality
As noted above, the IBM Watson Personality Insights service use Linguistic analytics A method of computational science that seeks to understand human language as it is spoken or written using algorithmic models. Linguistic analytics uses data a person generates through blogs, tweets, forum posts and email to generate a score relative to a sample population along a spectrum of cognitive and social characteristics. [See Glossary] to infer to an individual’s personality characteristics based on texts written by that person. This service is based on the Big Five model of personality Also known as the Five-Factor Model (FFM), this model proposes a taxonomy for personality traits divided in five dimensions or factors of personality: Openness to experience, Conscientiousness, Extraversion, Agreeableness, Neuroticism (or Emotional Stability). It is commonly applied to personality survey data to describe and correlate the human personality with other variables like job performance, life satisfaction and well-being, clinical disorders or entrepreneurial mind-set. This model is often referred to by the acronym OCEAN. [See Glossary], the most widely used model for generally describing how a person engages with the world.
4.2. Consumption Preferences
Based on the personality characteristics that are inferred from the text, the IBM Watson Personality Insights service return an indication of the author’s consumption preferences. Consumption preferences indicate the author’s likelihood to pursue different products, services, and activities. These variables have values between 0 and 1, being 0 unlikely and 1 very likely to prefer a certain product. Big Sister filtered items from these variables, excluding values equal to 0.5 because they are considered neutral.
4.3. Music Recommendation
One of the consumption preferences that are inferred by the IBM Watson Personality Insights service is Music Preferences, which indicate the interest of the author of the input text in different types or genres of music. The Big Sister app can function as a simple Recommender system A subclass of information filtering system that helps user(s) to select item(s) from a set of multiple items by inferring the “rating” or “preference” that the user would give to particular items. They are widely used in a variety of areas in platforms like Amazon, Spotify or YouTube. These systems follow complex algorithms to produce a filtering of items (books, songs, videos) and then predict items (or ratings for items) that the user may have an interest in, highlighting items that a user is more likely to interact with, share, or buy. These systems can use as inputs the user’s past behaviour (items previously purchased or selected and/or numerical ratings given to those items) and similar decisions made by other users (collaborative filtering), or the description or characteristics of each item (content-based filtering). Current recommender systems typically combine one or more approaches into a hybrid system. [See Glossary] by making a request to the API [See Glossary] of Spotify, asking for songs recommendations close to the genres that the author of the text was likely to prefer inferred by Personality Insights.
4.4. Comparing your results
Once you have performed several tests using different data sources or at different times, you can go to the Eye screen and click on this icon. “Your history” screen will open with all your results, differentiated by date and information source. By clicking on one of them, you can see your results in detail again. Returning to the “Your history” screen, by clicking on the three points and then in the option “Compare Sources”, you will be asked to select a range of dates.
Then you will see an interactive graph with all your results conducted in that period of time sorted by the source of information. So you can check if the inferred personalities differ between Twitter and Facebook or try different texts to see how their results vary. You can also see how your results change over time with a new tweet or post.
Tip: By clicking on the three points and then in the option “Clear history” you will delete ALL the tests performed and stored on your device, so please be careful.
Troubleshooting
If you are having problems while using the Big Sister app, you can write an email to the coordinator of the project that contacted you or directly write to the programmer of the app, Andre Simon (aasimon@uc.cl). Please include a screenshot of the error to help you determine and address the problem.
Privacy
All your texts and results will be stored only on your device. This data is processed for the non-commercial academic purposes of Algorithm Procedure or formula for solving a problem or accomplishing some end, based on conducting a sequence of specified actions. In computer systems, an algorithm is an instance of logic written in software to effectively produce or compute an output from a given input [See Input/output].ic Identities Project and will not be shared with third parties.