NETE Labs (@netelabs) at NET ESOLUTIONS CORPORATION (NETE®) is a pilot research unit developing expertise in digital technologies, research analytics and business intel. NETE Labs is a nascent project- the concept took shape around July/Aug 2016. NETE Labs is staffed by full-time employees of NETE(@netesolutions) who have common interests in challenging problems typically outside the immediate scope of their assigned responsibilities. TNETE Labs supports open source way and makes every effort to share its code and offer appropriate attribution. Being a corporation with obligations to its customers and entities that we lease data from, sharing code isn't always possible though. NETE Labs collaborates with academia and industry. Expressions of interest in collaborations are welcomed (email: netelabs@nete.com).
In March 2018, we submitted a manuscript describing ERNIE, a simple framework for acquiring, integrating, and analyzing citation data linked to other dimensions of research assessments. The manuscript will be posted on BioRxiv shortly. We used graph analytics in these studies using PostgreSQL, Neo4j, and Cytoscape.
In June 2017, we shared a preprint on BioRxiv of a study (being reviewed) we conducted in collaboration with folks from Elsevier and Gladstone Institutes. The paper was submitted to PLOS One but we withdrew it after not receiving reviewer comments over 90 days after submission, perhaps the process at PLOS One was overwhelmed by submissions. So we submitted the mansucript to Heliyon and it was published on Nov 15, 2017. Core data from this study are archived on Mendeley Data.
In April 2017, we shared Python code that we wrote to parse Web of Science (WoS) XML data. A version of this code has been successfully used to load the Web of Science Core Collection (around 64.5 million publications at last count plus millions of rows in related tables) into a PostgreSQL 9.6 database. WoS data is made available as a collection of XML files. The procedure used at NETE involves splitting each XML file into smaller XML files of 20,000 records each and then feeding the split XML files to the WoS XML parser, which extracts the elements of interest and loads them into 9 csv files. These are loaded into PostgreSQL tables using a loading script that is a part of the parser. These data are presently stored in the Azure cloud using a system architecture that consists of four Centos 7.2 VMs running 24x7 as well as an inducible 9-node Spark cluster that is run for a few days each month for large computing jobs. This architecture was developed for internal use by a federal agency under a contract to NETE Solutions. We've since made several enhancements to the code, and the latest version is deployed in the ERNIE project.
Several projects are currently active and involve collaborators from outside NETE. At present, they are mostly centered around the use of administrative, bibliographic, clinical trials, and other public records to enable studies of biomedical research in particular. The work of Williams et al is inspirational in this regard and and a related NETE Labs project was being archived on the Open Science Framework until we decided to try another approach to managing our collaborations. Relevant key words/phrases are 'bibliometrics', 'research evaluation', 'linked datasets', 'administrative records', 'patents', 'clinical trials', 'clinical guidelines', 'research awards', and 'researcher profiling'.