Please use this identifier to cite or link to this item:
https://www.um.edu.mt/library/oar/handle/123456789/91375
Title: | A DSL for business intelligence monitoring |
Authors: | Grech, Jean-Paul (2014) |
Keywords: | Domain-specific programming languages Information theory Electronic commerce Business enterprises -- Computer network resources |
Issue Date: | 2014 |
Citation: | Grech, J.-P. (2014). A DSL for business intelligence monitoring (Bachelor's dissertation). |
Abstract: | Companies are interested in certain patterns or events occurring in software systems since knowledge of these can benefit business. An example would be whether a campaign run by the company has garnered a certain level of popularity on its social platforms. Tools exposing these happenings and patterns are typically of the form of interactive dashboards which display a summary of the relevant events and update information as these events occur. The problem is that these solutions are typically inflexible such that the specifications describing the events are usually hard-coded and therefore leave the users limited freedom to express what to look out for. This project explores the possibility of enabling the users of a system to specify the patterns they find relevant through a language specifically designed for their domain as opposed to relying on a pre-packaged solution. The domain experimented with is that of Facebook page monitoring. In an attempt to find an alternative that overcomes these limitations, this project involved the creation of a language aimed at non-technical people. A compiler was created so that some of the lower-level specifications were generated from input obeying the language's grammar. The validity and range of the language has been examined through the use of interviews with people from industry. It was found that the language allowed for added flexibility but did not encompass the whole domain. The language's ease of use was assessed by asking a non-technical audience to explain some phrases written using the language. This revealed that the language was not as simple as intended but still comprehensible nonetheless. We concluded that a language is viable for the specification of patterns in the context used. However, to be fully effective, those responsible for the language's creation must have excellent knowledge of the domain and topics that apply to it. For example in our case not only would experience in social media management be necessary, but also a good understanding of marketing. A particular kind of technology was used to specify and check the patterns users are interested in observing. This technology specifies properties and verifies whether they are upheld. We presume that the simplification of property management granted in this project might serve as added motivation in applying these techniques to a larger part of the Business Intelligence field. The achievements of this project show that a domain-specific language is a viable solution in allowing the creation of necessary properties related to Facebook pages for the domain of Business Intelligence. Given that other areas of this domain can be suitably broken down, it could be argued that the techniques used in this project can also be applied to these. Perhaps other social platforms such as Twitter and Google+ can be incorporated to give a richer analysis of a business's status. |
Description: | Bachelor of Science (Computing and Business) |
URI: | https://www.um.edu.mt/library/oar/handle/123456789/91375 |
Appears in Collections: | Dissertations - FacICT - 2014 Dissertations - FacICTCS - 2010-2015 |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
B.SC.(HONS)ICT_Grech_Jean-Paul_2014.PDF Restricted Access | 4.95 MB | Adobe PDF | View/Open Request a copy |
Items in OAR@UM are protected by copyright, with all rights reserved, unless otherwise indicated.