Information overload has consumed us for more than a decade, spawning a generation of social, collaboration, search, and analytics solutions to help the knowledge worker more effectively consume content. As we enter the Cognitive Era, data overload is the new challenge and the knowledge worker now needs help to consume (and derive insight from) data. Despite having increasingly powerful analytics tools, the volume of data combined with limited analytics skills, is impacting our ability to be successful. Insight discovery requires collaboration and active engagement from across the business; from developer to data architect, analyst to data scientist, executive to business user. We need personalized and contextualized ways to find data, to share insight and expertise, to build and sustain communities of practice, and to more effectively apply data to business problems.
Is collaborative analytics the next frontier for social and collaboration systems? I believe the answer is Yes, and in this blog post I’d like to explain why. I’d like to share with you my current thinking on this topic, and not too surprising for those who know me, I’m specifically looking at collaboration as a way to capture the digital breadcrumbs that make up the analytics process. I’m interested in the broader concept of “interaction data” and how we can use these interaction breadcrumbs to gain greater insight into the analytics process and to ultimately provide a more cognitive, contextualized, and personalized data science experience. This new cognitive experience would deliver value to all participants in the analytics lifecycle (data scientist, business analyst, executive, business user, …) by streamlining and creating greater transparency around the analytics process from data creation to consumption, insight generation to decision making, and everything in between.
Today the Data Scientist is isolated which is not a good thing for either the data scientist, the business, or anyone else who may be involved in the analytics process, directly or otherwise. For a successful analytics project, people with many different backgrounds need to collaborate; to have group discussions in context. They need to be able to see what their colleagues are doing, to be able to find the data they need, when they need it (in context); and sometimes what they need to find is people, not data.
To address this isolation challenge, which in turns creates a challenge around transparency, provenance, and potentially governance, we need to embed collaboration into the analytics process in order to connect the data scientist with his team members, and with the broader analytics community for more effective skills and knowledge sharing. This sharing touches on the entire life-cycle of data, from crowd sourcing data semantics to sharing information about data consumption, quality, transformation, derived insights, and ultimately efficacy.
This collaboration experience can come in many guises, however I see the activity stream as being a key component. The stream allows you to have a single view of the project and supports social services, such as comment, tag, like, share, chat, …. It allows you to define projects and teams, and to integrate with 3rd party analytics tools, such as Watson Analytics, where it can become an integration hub for the analytics project and team. It captures and persists all interactions during the project, providing context and maintaining a provenance trail.
Now this is the fun part ;-)
All interactions are modeled in the Contextual Knowledge Graph; human-human (collaboration), human-data (consumption/interaction), data-data (transformation). Once this data is captured and modeled, it allows us to apply analytics that will deliver a personalized and contextualized cognitive user experience.
Circling back to the original question “Is collaborative analytics the next frontier for social and collaboration systems?” I hope I’ve managed to convince you that the answer is indeed a resounding Yes! If you are still not convinced, then here are some questions to mull over…
- Do you have a distributed analytics community?
- Are you looking to leverage data, skills, and insights across business silos?
- Is analytics upskilling a priority, and challenge, for your business?
- Is analytics provenance important to you?
- Are you looking for greater transparency within, and understanding of, your analytics projects?
- Do you have a culture of collaboration within your business in general? A culture that would make collaborative analytics possible
If you answer Yes to any of these questions, then it’s likely that collaborative analytics should be on your analytics roadmap.