AI Coding
Categorize and code your survey responses with AI
Introduction
Aftercare takes a relationship based approach to / coding survey responses. We believe that themes and sub-themes are best defined by the relationships between them.
Aftercare uses a tree of themes and sub-themes to code survey responses and will place the response in the most specific category.
Tree Maintenance
As new responses come in, get coded, and begin to branch out, the coding tree will quickly grow. Aftercare provides a few knobs to tune how much you want to prune the tree.
Take the wheel
Aftercare will auto-magically create new themes and sub-themes for you as responses are coded. Let your tree grow organically.
Bonsai Trees
Mould the shape you want your tree to take. Lock down the branches you want to keep and let others get auto-generated.
Manual
Manually create the themes and sub-themes and lock them down. Aftercare will not create new themes or sub-themes.
Tree Growth
Aftercare will provide sane default to grow your tree, but we also support additional growth factors. Tree growth is determined by a few parameters and tuning these will change the shape of your tree.
| property | description |
|---|---|
max levels | how tall your tree is |
max codes per level | how wide each branch gets |
min codes per response | when create new themes, how many responses are required to do so? |
Climbing the Tree
When a response is coded, Aftercare will start climbing the tree to find the right codes to visit. At every branching path, there are a lot of decisions to make. This section will give an idea of how we traverse the tree, but it is not exhaustive to the logic that goes into making these decisions.
Data Quality
Is the response gibberish? Are there multiple components to the response that could be interpreted as different themes?
Evaluate
Does this response match any of the existing themes? Should we create a new theme? Can we create a new theme?
Repeat
Determine if we should repeat the process for each branch.