toggle quoted messageShow quoted text
I’ve been following your Amazon book mapping activities for as long as you’ve been doing them. They do paint a persistent and persuasive picture. However the other day, I wondered how you would know whether the data behind the maps is contaminated - do you have good insight into how Amazon's algorithm works for its suggestions? For example, do you know whether it is simply driven by “people who bought also bought” or are the results further tuned by topic categorisation putting “like” books together - hence accentuating the separation effect?
Do you have any insight into this?
>> Valdis, do you have any ideas about data collection in this environment to build the SNA off? I wanted to get access to the phone records but got a hard NO due to union issues. Also, given the highly transactional nature of emergency management, most of those calls will be to communicate or hand-over incident information, not pass on knowledge or solve problems.
There are ways to collect data and keep it private/anonymous. It all depends on what you have. But if an organization does not want to hand over data, then there is not much you can do.
Think about whether there are any “proxy networks” available that will give you insight into the organization without revealing anyone’s identity. Like the attached network map … it is based on actual Amazon purchase data, but none of their customer's identities are revealed. Yet we see emergent group behavior. How can this metaphor be used in your situation?