The aim of the project is to characterize the behavior of entrepreneurs who access social networks , and how they use social media to generate new business.
Consider using NodeXL and/or Pajek. I think you may find my book on analyzing the seven layers of social media data useful in this regard: https://7layersanalytics.com/introduction-to-the-book/
Consider using NodeXL and/or Pajek. I think you may find my book on analyzing the seven layers of social media data useful in this regard: https://7layersanalytics.com/introduction-to-the-book/
Big Data analysis requires appropriate tools. If you want to perform some network analysis, I suggest to use Giraph or Spark GraphX, which include many ready-to-use algorithms.
I find NodeXL works well identifying "two-dimensional" snapshot of a given set of interactions within a social network. However, it has trouble identify the movement of conversation based on a temporal scale. I haven't found a more "three-dimensional" analysis to this problematic.
SNAP (http://snap.stanford.edu/snap/download.html) is developed at Stanford and is delivered to analyze the large networks the university collected (http://snap.stanford.edu/data/index.html). It is delivered as a C++ app, but you can find the website a Python Wrapper.
Otherwise, if you design your own algorithm, you can use sparse matrices format. Baiscally, all social networks are sparse (a lot of "0" in the adjacency matrix), and coding your algorithm using this kind of data structure will fasten a lot the computations. This kind of data structure is used in the Louvain algorithm (https://sites.google.com/site/findcommunities/) and it exists in Python (http://docs.scipy.org/doc/scipy/reference/sparse.html).