Issue link: https://maltatoday.uberflip.com/i/1088108
16 INNOVATION maltatoday | SUNDAY • 3 MARCH 2019 LARGE volumes of data are be- coming increasingly available to organisations through open data sources and this information is being used to improve decision making. Due to the ever-increas- ing availability of data over re- cent years, acquiring data science and analytics skills has become important to the industry. Data science and analytics techniques can help in providing a better understanding of stakeholders' behaviour and trends. Govern- ments can benefit from data ana- lytics to better understand the impact of the policies that are put in place. As companies, organisa- tions and governments begin to rely more and more on data for decision making, the demand for data scientists will grow rapidly. About the collaboration To address the high demand for technical skills and support the rapid changes within the academic curriculum, MITA's Programme Management De- partment (PMD) and Digital Outreach Team collaborated with the University of Malta's Faculty of ICT through a Frame- work Agreement back in 2017, to carry out several activities with students. One of the many ini- tiatives listed in the Framework Agreement was the organisation of an Academia-Government Datathon. This was a semester- long activity during the 2018/19 academic year; which culminat- ed in February 2019, during an event where students presented their projects. Students reading for the Master of Science degree in Artificial In- telligence within the University of Malta were given a simulated dataset within the context of Taxation, that was supported by the Commissioner for Revenue. This dataset was provided as part of the Mining and Visualising Large Scale Data credit, and con- sisted of over 163 million data items. The main goal of this pro- ject was to leverage technologies and tools from Malta's Hybrid Cloud enabling infrastructure, in the areas of Visualisation, Text Mining and Graph Mining and use them to provide potential so- lutions to problems in Taxation and Financial Policy. The main benefits of this col- laboration were primarily to help students improve their skills in data analytics/visualisation of large-scale datasets, the possibil- ity to develop proof-of-concepts within the mentioned areas, and the exposure to use the lat- est cloud technologies, tools and services. A pitching event took place at the MITA Data Centre in No- vember 2018. Each student pre- sented the ideas for the problem solution. Students were given feedback by academics and practitioners during discussions on the intended use of the pro- vided dataset and technologies. To further support the students in understanding the tools avail- able on the new IT infrastruc- ture, MITA delivered a lecture on the Government's Hybrid Cloud enabling infrastructure and the advantages of using this technology. The lecture then delved into practical content on cloud technologies and data an- alytics. Students were also given the opportunity to make use of the MITA Emerging Technolo- gies Lab by utilising Mixed real- ity devices for visualisation pur- poses. On Friday, 8 February 2019, the students presented their re- sults and demoed their solutions. Presentations on the visualisa- tions and data analytics in the form of dashboards and similar concepts were made. The Per- manent Secretary for the Minis- try for Finance (Merger and Ad- ministration), Joseph Caruana was also present for this event. A judging panel made up of two MITA experts; Marvin Zammit and Keith Cortis, and the stu- dents' academic lecturer Charlie Abela, selected the most creative and innovative solution. MITA's Hybrid Cloud Malta's Hybrid Cloud pro- vides a cloud computing envi- ronment that combines private (on-premise) and public cloud (off premise) services based on the Microsoft Azure platform. The students were given access to the public part of the Hybrid Cloud, which is serviced from the Microsoft Azure public cloud, and thus offers all cloud computing services provided by the Azure Service catalog. The students were encouraged to use the cloud services which were re- served specifically for this study unit. Technologies used Given that the goal of the project was to address a de- fined problem through a solu- tion that allows for increased value and insights by leveraging techniques and technologies in the areas of Visualisation, Text Mining and Graph Mining, the relevant cloud services and/or technologies were identified for use by the students. The follow- ing are the main ones identified: Power BI for Visualisation; Ma- chine Learning Studio and Text Analytics service for Text Min- ing and Cosmos DB for Visu- alisation, Storage, and/or Graph Mining. Other Infrastructure (IaaS), Platform (PaaS) and Software (SaaS) cloud services, such as Virtual Machines (e.g., Neo4j Enterprise), Virtual Networks and SQL Database, were used within the projects as deemed necessary. This includes exter- nal services that are published by third party software vendors in the Azure Marketplace, such as the Neo4j Graph Platform, which is used for graph storage, mining and visualisation. About the students' projects Several innovative solutions were developed by the students in the areas of Taxation and Fi- nancial Policy to address chal- lenge and visualise the analytical results from the dataset in an in- teresting and useful manner. Identifying tax evasion and fraud were objectives of solu- tions presented by some of the teams. Other projects focused on the text mining aspect, used techniques, such as pre-process- ing, key phrase extraction and sentiment analysis, in their solu- tions. Such projects ranged from identifying the most promising industry area to invest in, and visualisation of economic trends over time, to predicting the like- lihood of bankruptcy, analysing the Maltese economy during recession periods, extracting knowledge on the economic situation of particular sectors, and the impact of Government Budgets on the economy. More- over, a graph mining-based so- lution predicted the lifecycle movement of companies and their industries. Besides the simulated Taxation dataset, which was of a financial nature, other external data sourc- es, such as news articles, financial lexicons and domain-specific dic- tionaries, were used by some of the developed solutions. Through this collaboration, the students demonstrated their creative skills and understand- ing into utilising Visualisation, Graph Mining and Text Mining tools and techniques on the data provided. Feedback received The students' lecturer Charlie Datathon: data is the new currency A collaboration between MITA and the Department of Artificial Intelligence within the University of Malta Keith Cortis and Daniela Chetcuti Programme Management Department