For those of us that work in the world of collaboration, social business, or whatever you want to call it, we spend a lot of our time focused on facilitating collaboration around content; helping people share their ideas, build and sustain community, run projects, find people, experts, content, … all the things that are necessary to function in an interconnected and frequently distributed enterprise.
However what about the data scientist whose day is predominately focused on data and not content, do they get any of our attention? What have we done recently to help the data scientist, or indeed the data or business analyst, collaborate around data, where ideas are articulated in numbers, charts, visualizations, algorithms, and not necessarily content.
Last week I was at the IBM Connect conference in Orlando and spent some time with one of the IBM Research projects out of the Almaden Lab which it demonstrated something which they referred to as Collaborative Analytics, where the data scientist could perform their analytics in an environment which supported light-weight collaboration, context building, and provenance tracking wrapped around their data manipulation and analysis. While Notebook technology provides a great way of sharing analysis between people, this system allowed interactive collaboration between multiple parties in-situ as the analysis was being performed.
And even better, the entire system was built natively on an enterprise graph (using my favorite graph platform; Titan & Tinkerpop) which captured the social, semantic, and physical connections between all actors, assets, and interactions within the analytics process – people, data, transformation, analysis, visualization, content, commentary, tags, … – and then used the enterprise graph to provide recommendations that allowed the data scientist to identify interesting (and highly relevant) new data sets, analysis results, and (most importantly) people spread far and wide across the enterprise.
I will definitely write a more detailed post on this solution, but for now I wanted to finish with the question that opened the post…
Have we been neglecting the Data Scientist over the last few years? Considering how much data is spread across the enterprise, the growing demand to leverage this data for business value, the generally limited data science skills within the enterprise, and the increasingly powerful analytics platforms that allow this data to be consumed and analyzed in new ways, should we be doing more to help people across the enterprise more effectively find data, expertise, analysis results, and apply them to their business problems.
Is collaborative analytics the next frontier for social and collaboration systems?