Resumen de Big data and social networks in data envelopment analysis
The present study contributes to the body of knowledge in the filed of efficiency and Data Envelopment analysis. We introduce Social Network Analysis (SNA) and big data into DEA, in order to develop new methods and features, and draw out practical applications to solve real-world problems, e.g. efficiency analysis of peer-production organizations. Social coding in Free, libre and open source software (FLOSS) organizations is an emerging and rapidly growing phenomena of peer-production that has gained a lot of attention from scholars, however, little is know about its efficiency.
Three core chapters of the present study are (1) DEA and stepwise benchmarking: The introduction SNA-DEA–a network-based step-wise benchmarking approach that takes into account input similarities and efficiency difference to be improved in each benchmarking step. (2) DEA and big data: The introduction of BigDEA–an efficient method that enables DEA efficiency analysis in the big data context. (3) Temporal BigDEA–an extension of panel data DEA and BigDEA that enables long-term efficiency analysis in the big data context.
In the first chapter we introduce SNA-DEA then we apply it to a network of bank branches and discuss the practical implications. In the second chapter we discuss BigDEA and utilize it in order to analyze the efficiency of more than 500,000 peer-production DMUs. We also define a measure called merge ratio (the ratio of merges i.g accepted pull requests by project maintainers over the requested pulls by open source contributors) which predicts the efficiency score of organizations very accurately. In the third chapter we focus on temporal DEA which deals with the long-term VRS, CRS and scale efficiency analysis in the big data context. Then we apply the method to a large panel data of software peer-production organizations with more than 46,000 DMUs for the first quarter of 2016 to the last quarter of 2017, in order to analyze the trends of efficiency evolution over time.
