It’s so funny to hear this because I think in an alternative way about why Design research and data science work together I know that observational data, ethnography methods (not so much interviewing) can find insights that can never be found in data because data is not truth. And so ethnographic methods help bring out the hidden codes, ritual, cultural stuff that often quant data misses. Especially with small sample sizes.

I often find when I using tracking data and people know their being track it engineers a larger shift in behavior change than observation at least in the beginning. It takes a couple of weeks for people to forget someone is monitoring their data. Also self-quant is different. The behavior isn’t as pronounced when the participants is the one monitoring her data. It’s for this reason when I use tracking data I’m always on the look out for anomalies and outliers rather than trends and patterns because I need to test the truthiness of the data I’m getting.

I love data science methods for prototyping, especially sense/act/learn validation and inspiration for intelligent products.

But when it comes to exploratory research I tend to keep them compartmentalized trying to see what story the numbers are telling me and what story the ethnographic data is telling me and finding the seams. In the seams are usually design opportunities.