This November we spoke about NLP and Text Mining techniques used in working with open Government data in a practical way.
The topic was to “Reveal hidden patterns in public data” with two great experienced speakers – Anton Gerunov and Yasen Kiprov.
Anton Gerunov has served as e-governance lead for the Bulgarian government (2014-2017) and headed the open data initiative in the country. This led to the opening of over 6,000 data sources on the open data portal, making Bulgaria one of the leaders in this field (16th place in the Open Data Index in 2016, up from 51). He has had experience in the private sector as international business consultant and project manager, and in the public sector as innovation and analytics advisor to the Prime Minister. Anton Gerunov holds a PhD in Economics and is now an Associate Professor of Risk Management at Sofia University and COO at LogSentinel – the blockchain-enabled secure data audit trail.
He spoke about “Public data – benefits review and technical challenges“
Anton presented a bird’s eye view of public procurement and outline main trends and possible uses of this dataset. The presentation also put public procurement in the context of inequality analysis and social network analysis (SNA), using R and Gephi.
Our second speaker Yasen Kiprov has been doing Natural Language Processing for 8 years, mostly in java and python. He has experience in several startups working with semantic technologies and NLP and recently became A.I. team lead at Siteground. Yasen was part of the winning team Chereshka at the last Datathon 2017.
He presented us how “Data reveals corruption practices – case from Datathon 2017“
His team, Chereshka, combined for two days the Trade Register and Public Tenders data. Yasen told us how linked data helped the team integrate and query the two sources. He also showed some interesting initial findings, including people who participate both in the tender request and in the management of the selected bidder.