Skip to main content

Business Intelligence

Business Intelligence:

For efficient operation business or corporate entities in general (including governments) need to make best use of the information they hold in their organisation and information outside their organisation that affects the business. We mine textual information from enterprise documents, social media and other sources to support businesses and assess social licence to operate. We also look at formal information flows in business and try to understand the socio-technical challenges that businesses face.
 

Current projects:

Open innovation (Pathways to market industrial transformation research hub):

Firms need to continuously innovate to maintain or enhance their competitive advantage. Technological advances are driving the proliferation of increasingly complex knowledge and firms no longer can solely rely on their internal knowledge assets for innovation. A growing number of firms are embracing open innovation as a competitive strategy. Open innovation is a distributed innovation process based on purposively managed knowledge flows to accelerate internal innovation and expand markets for external use of innovation. Data61 is exploring how social networks affect absorptive capacity in open innovation collaborations. A major goal is to gain insight into personal and organisational factors influencing boundary-spanning knowledge sharing, idea generation, idea realisation, and trust networks, and how these different networks relate to each other. A secondary goal is to discover how tacit knowledge and power relationships influence intra- and inter-organisational learning processes that underpin absorptive capacity.

The longitudinal spine of government functions:

For many years the action of government such as the money it spends and the records it keeps have been reported based on government’s function. That is how it applies itself to tasks rather than the hierarchical organisation agencies. The government can organise its actual structure as it sees fit to carry out the tasks. Under the National Innovation and Science Agenda’s Platform for Open Data, Data61, with the Department of Finance and the National Archives of Australia and CSIRO Land and Water, is bringing to function definition data from across government. Different parts of government uses different language to describe what it does. The spine will bring together 10 years of how Government has described what it does and create linkages between them and over time. Other datasets, such as money spent, can be then bound to the spine to perform analysis. For this information the expenditure on functions can be tracked, records located and the movement of corporate knowledge understood. This work is also applicable to and large organisation the under goes mergers and restructures.

More information is available here.

Network:

The most valuable part of an organisation is its people. However finding people in an organisation with the right skills to carry out a task is often down to who you know. Network leverages the artefacts employee create in an organisation to help identify them, their skills and relationships. For example many employees will have a profile page, but it may not be up to date. They may also create and file documents, be assigned to projects and be part of organisational structure. By mining this information from corporate APIs Network is able to build and up to date composite of people. In Network this information is then indexed and searched via custom search engine and displayed in a way which highlights the linkages between people and other things. Having this information available is a step towards finding the right people to create diversity, high performing, mission focused teams.

State Library of NSW Social Media Archive:

The State Library of NSW licences our Vizie software which provides the ability to archive and analyse publicly available social media data from multiple social media platforms. The tools help with the manual and time-consuming task of query formulation, providing assistance with ambiguous queries and offering a federated search interface to multiple social media search engines. The data collected is analysed and dashboards are designed for the ongoing curation of the library staff. The project uses other social media analytics capabilities developed within KDM, such as the emotion analysis modules. Through this project, the KDM group have developed a general social media processing pipeline that is in use in other KDM projects, such as the Social Licence to Operate project and the Syndromic Surveillance project.

In 2017, aggregate statistics from the archive was made publicly viewable and launched by the library one International Digital Preservation day.

More information is available here.
The archive: https://socialmediaarchive.sl.nsw.gov.au

Social License to Operate:

Using KDM’s in-house social media processing platform, the group has assembled a prototype to examine the Social Licence to Operate of different mining companies within Australia, examining this social licence. This prototype uses text analytics combined with sentiment/emotion analysis to provide a statistical model of social licence to operate. The demonstrator also uses stance detection methods to identify public opinions on various key dimensions related to the mining sector.
 

Completed projects:

Edible Sensors Challenge:

Edible sensors are sensing solutions that are fit for human consumption and capable of monitoring the state of a food product to determine provenance/traceability, food quality, freshness or related nutritional value. The purpose of this research was to explore (a) potential uses for edible food sensors, (b) what features actors in the food supply chain would seek from edible food sensors and (c) who the key stakeholder groups are that would be interested in, or affected by the use of edible food sensors.

Privacy in the FSC Challenge:

This project aimed to inform the development of technology platforms which deliver connected and visible food supply chains. The purpose of the research was to gain an understanding of the perceived opportunities and risks associated with sharing data across food supply chains.

Data Attitudes and Practices Survey (CSIRO’s Data Governance Initiative):


This work distributed a CSIRO wide survey to employees with the aim of understanding the organisational, sector-specific, disciplinary and individual factors influencing research data sharing.
 
 
 
 
 
 

Understanding the food supply chain from the consumers’ perspectives:



In this project we investigated consumer perception of and trust in food and the food supply chain, via mining of social media data.
 
 
 
 
 
 

Commemorating the Centenary of Armistice on Social Media:

During 2018, the Australian War Memorial commemorated the Centenary of Armistice over a five-week period from 5 October to Remembrance Day, 11 November. The theme of commemorations was ‘Honour Their Spirit’, and a number of public engagement activities were held, culminating with the Remembrance Day National Ceremony. In this project we processed social media content from Twitter and Instagram in the lead up to this event to associate the messages to one or more of the six emotions of gratitude, humility, pride, respect, sadness and discontent. This formed the basis of the Emotion Poppy, an interactive, web-based visualisation displaying the changing sentiment or mood expressed in the real-time social media posts discussing topics relating to Remembrance Day.

Value netchain analysis (Australian Centre for International Agricultural Research [ACIAR] – Beef value networks in South and Central Highlands of Vietnam):

Value netchain analysis is a process of understanding, mapping and analysing sets of roles, interactions and relationships that generate economic (or social) value to a member of a social network. It explores how the tangible and intangible resources within a network eventuate into value.

Information Extraction from PDFs: An example from the Retail Leasing Sector with LeaseInfo:

Accurait®’s achievement is a unique AI solution that reads the scanned text of lease agreements to a high level of accuracy and also extracts and interprets the meaning content in complex legal agreements. Responding to the new IFRS16 international accounting standard, Accurait® is the only automated solution in the Australasian market that can accurately and rapidly extract ‘fine-grained’ information from commercial leases to a level that is directly relevant to compliance with this standard. For the user, Accurait® supports an end-to-end process. It is thus a genuine disruptor in the global $2.8 trillion dollar leasing industry. This massive undertaking was only possible through the creative realisation that LeaseInfoTM’s historical data and extensive industry knowledge could be coupled with CSIRO’s Data61 natural language processing expertise (within KDM). This partnership allowed a rapid generation of the data sets (so rarely available) required for machine learning, thus producing an agile response to market interests.

Accurait® is a significant advance that value adds above traditional Optical Character Recognition tools by coupling these with a new information extraction pipeline tailored for contract leases, particularly those that are scans of originals, captured as a PDF, rather than “born-digital” documents . To develop this unique solution, LeaseInfoTM and CSIRO’s Data61 ‘trained’ the AI on millions of contract pages such that the extraction method can be applied to the different legal jurisdictions of Australia with their varying document formats. This massive undertaking was only possible with the unique collaborative effort: LeaseInfoTM, contributed the requisite historical data and its extensive industry knowledge (including user requirements), and CSIRO’s Data61, using methods developed in-house and the CSIRO High Performance Computing Cluster, rapidly generated the data sets required.

More information is available here.
The product: https://www.accurait.com.au/
 

Relevant expertise:

  • Survey design and analysis
  • Mixed methods research
  • Social network analysis
  • Exponential random graph modelling
  • Multilayer social network analysis