distributed-ethnography
Distributed ethnography is a method of understanding a "community’s belief systems through their own narratives" (K Mausch, 2018). It tackles the idea that it's very difficult to use technology to understand what someone means by having the storyteller quickly explain the meaning of their story across a set of dimensions. These dimensions can then be used as metadata. Technology can then use this metadata to aggregate and slice stories from hundreds or thousands of storytellers, and provide the end consumer of this information a way of understanding trends and then diving into the individual stories.
I first came across distributed ethnography through a speech given by Dave Snowden (D Snowden, 2016). In that speech, Dave makes a very strong case for the weaknesses in traditional survey-based, and presented his approach of having survey respondents tell a story and then rank the meaning over their stories across sets of three positive qualities. This his argues -- and my further reading seems to support this argument -- can significantly reduce bias in the results.
However, the only tool that I've seen that supports running distributed ethnography is closed-source and maintained by Dave Snowden and his company The Cynefin Co. This opens up distributed ethnography to some criticism that it is just the same old survey-based research consulting under some excellent marketing-speak, but I'm not convinced. In the hopes that I'll get an opportunity to put some of this theory to the test, this repo serves as an open-source tool to experiment with distributed ethnography.
Setup
The goal is everything in this repo can be accomplished via make.
Note: This repo was written on MacOS, but any Unix-based OS should work.
Toolchain
To run this service, you will need the following programs on a Unix-based operating system:
| Tool | Notes |
|---|---|
| sqlite3 | Simple data storage solution for the application. Converting to a backend like Postgress is possible. |
| yq | Used to parse TOML configuration files to set local and testing environments. |
| make | Used to hide the headache of managing python environments. |
| python3 | This application is written in python 3.12, but likely works for versions >= 3.9. |
Install these from your respective package manager.
Database
To setup the local sqlite3 database, pick a location to house the database and update /config/local.toml to point SQLITE_FILE to this location.
Then, run make setup-db.
References:
- Mausch, K., Harris, D., Heather, E., Jones, E., Yim, J., & Hauser, M. (2018). Households’ aspirations for rural development through agriculture. Outlook on Agriculture, 47(2), 108-115. https://doi.org/10.1177/0030727018766940
- Snowden, D., (2016) How leaders change culture through small actions [Conference presentation]. Bangor University. https://www.youtube.com/watch?v=MsLmjoAp_Dg