The data scientist should read all agreements and policies of the specific social media site before engaging in data mining for sentiment analysis. The agreements and policies are carefully written by social media companies to protect users, the social company, and the developers. On Twitter’s Developers Agreement under “VII. Other Important Terms” there is a guide on what information is available and how it should not be used (Twitter Developer, May 25, 2018). By acknowledging the agreement developers who have access to Twitter data cannot use the data to give public agencies the information for surveillance, investigating, or tracking users and their content (Twitter Developer, May 25, 2018, VII, section 1). A data scientist should be made aware of the use of the data analyzed. The agreement stresses that any collected data not be used for investigation, surveillance, analysis, or research that discriminates against individuals or groups and must adhere to the users’ “reasonable expectation of privacy.”
(Twitter Developer, May 25, 2018, VII, section 3). Prohibited is any targeting of individuals health, financial status, political stand, beliefs, ethnics, religious, sex, etc. under the personal information that is prohibited by law (Twitter Developer, May 25, 2018, VII, section 4). Under Twitter’s Developer Policy Part C is “Respect Users’ Control and Privacy” (Twitter Developer, November 3, 2017). This section is specific to developers who have apps that a user is maybe interacting with their Twitter account(s). Twitter says that full disclosure of what it uses the information for and how it handles the information with the ability for the user to opt-out, delete and altogether remove all data, right to privacy, confidentiality, and disclosing what is published by the user (Twitter Developer, November 3, 2017). The data scientist needs to read all the terms before doing anything with mined data from anywhere. Is the project internal research for some broader project scope, or is the sentiment analysis the subject of a published work by the data scientist? Both cases shall adhere to the agreement that Twitter has provided to get access to the site’s data. Sentiment analysis is fine if it doesn’t target protected groups and/or the identities of users or provide any other information that is prohibited in the written agreement/policies.
References
Twitter Developer. (May 25, 2018). Developer agreement. Retrieved from: https://developer.twitter.com/en/developer-terms/agreement
Twitter Developer. (November 3, 2017). Developer policy. Retrieved from: https://developer.twitter.com/en/developer-terms/policy