CASM Consulting is committed to ensuring that all and any data used to conduct our work is handled to high ethical standards. To this end, the following principles are observed in each of the projects we undertake:
Collection of data
Our work often involves analysing data which is in the public domain and available online, including data shared on social media and available on web pages. Any such data is collected according to the following principles:
- CASM Consulting does not and will not collect or process data which has not been made publicly available, unless we have express permission from the subjects of that data to do so.
- Data collected from social media platforms will always be collected through that platform’s official API, where such an API is available, and all applicable user and developer agreements will always be followed in collecting, storing and processing this data.
- Data collected from websites, such as online forums, will only be collected from forums accessible to the general public (i.e to which no account is required to view) and will not be collected from pages which restrict or forbid automated collection
- Some projects undertaken for specific clients may involve the processing of data held by that client. Where this is the case, this data will only be processed where the client has been given specific consent to share it.
CASM uses the data we collect to study broad trends; for example, the changing national opinion on the effectiveness of policing. Data is not and will not be used to study groups or individuals who have a reasonable expectation of privacy, to conduct surveillance or to enable individuals to be profiled and marketed to. Furthermore, all data relating to individuals is always anonymised before publishing or sharing with partners – for details of how we do this, see ‘publishing and sharing data’ below.
We conduct this analysis using Method52, a platform developed and owned by CASM Consulting. Method52 allows us to build applications which are sensitive to the ethical requirements of each project.
Analysis conducted using Method52 might include.
- Training algorithms to analyse the free text of shared posts and web pages to work out which topics and themes are being discussed within them
- Use of algorithms to ‘cluster’ large numbers of posts and discover common phrases and themes which occur within a dataset.
- The use of existing data within a post – e.g. the timezone in which it was created – to infer geographic characteristics such as the country or city from which a post is likely to have been sent.
- In general, CASM Consulting does not collect or process the personal data - e.g. the names, addresses or telephone numbers - of individuals. Where collecting this data is unavoidable, for example, where social media account names are returned from a platform’s API, these details are never published or shared with third parties.
- All data collected is stored securely on servers accessible only to named researchers working on each project, and a small and controlled group of technical administrators.
- Data is encrypted in transit using HTTPS
- All data collected by CASM Consulting can be verifiably erased, and we do not store data for any longer than each project requires.
Publishing and data sharing
As a core principle of our work, CASM Consulting is committed to protecting the privacy of individuals who appear in our datasets. To this end, the following precautions are taken when analysis is shared or published:
- Analysis presented to any third party is always presented at an aggregate level, showing e.g. trends applying to a population as a whole, rather than the behaviour of individual accounts
- Usernames and given names are never shared or published within reports, with the exception of users who are not likely to have a reasonable expectation for their public posts to remain unpublished (e.g members of parliament etc.)
- Any quoted posts or excerpts of free text are bowdlerised – i.e the meaning is preserved but the wording altered – to prevent a user’s being retroactively identified through searching for the text of the post.
We’re happy to answer questions on how we process data. Please use the contact form below, or the mailto icon at the bottom right of our page.