This repository was created to continue the work from the Microsoft Financial Services Autism Hackathon for use case #3. The goal was to see what information could be drawn from a pool of de-personalized ABA related data as a demonstration of how modern tools could be applied to a relatively unexplored business problem. Success would not mean we need a specific answer, but could be that there is a better articulation of the challenges so that partners in the industry could have a clear roadmap of how they could adapt their data collection so that it could be used to inform the overall community.
This use case will analyze anonymized therapy data from a repository of Applied Behavior Analysis sessions and cases to draw insights which can be used by individual families to help understand the effectiveness of their programs. Success criteria for programs are boosting skill acquisition or reduction of interfering behavior
There is a long-held belief that each individual case of autism is so unique that comparisons cannot be drawn to other cases. However, we know that there are certain characteristics which define autism (required for diagnosis). To understand the progress of their child, families need the ability to perform comparisons and reference benchmarks. Can we use the data to help them accomplish this?
Can we find commonalities among cases to create segments and find benchmarks based on looking at the data alone?
Can we determine patterns in skill acquisition which can help segment and benchmark?
Can we identify patterns that would allow grouping of cases?
Can we identify programs that are successful for these identified groupings?
If there is a larger group and specific interests we could also try to utilize visual aids, such as a Surface Hub to help visualize and manipulate the data.