Data Management Strategy
ICT 4005 – Technical Foundations by Stephen Barnes
January 21, 2018
Why manage data?
Over the last few years, development in new technologies like smart devices, online courses, flexible classrooms, etc. are completely rebuilding the approach towards learning and teaching. Many educational institutions are embracing these new technologies and methods to meet today’s needs. The universities need their data to improve as they are now operating in a progressively complex and competitive environment. These days, the sources of data such as educational content, enrollment, scholarships, etc. are enormous. Managing, analyzing and reporting across departments and/or campuses is becoming more challenging.
To make better decisions organizations have been using data analysis for many years. According to IBM, 80% of the data generated by organizations is unstructured and can be in various formats like text documents, video, audio, diagrams, images and multimedia documents (Schneider, 2016). The university’s data volume is increasing every year, and most of it is unstructured and scattered, hence making it difficult to retrieve or consolidate. New technologies have emerged that have made possible to collect and store these complex data from multiple sources. There are analytical tools available which can convert these unstructured big data into meaningful data.
Machine learning is the current trend in today’s tech world. Growing numbers of systems and tools take advantage of it for big data analysis. It is a method of data management, analytics, data visualization and reporting e.g. – SAS. Tools for data storage and management such as Hadoop, Cloudera etc., data mining tools like IBM’s SPSS Modeler, data analysis tools Qubole, Statwing, data visualization tools Tableau, Silk and much more are becoming more advanced (Import.io, 2017).
Data Management Policies
I believe by implementing a data management policy, which establishes data governance roles and rules, the University will be well-positioned to implement and improve its data strategy (framework adapted from Flory and Walker, 2015). The following are the steps involved –
1) Data Governance – A formal data governance structure will help resolve various data issues currently present in the University data management policy. “A robust data governance framework will provide the structure and institutional oversight necessary to establish a culture of data fluency across the institution” (Flory and Walker, 2015). This can be implemented by following these steps –
a) Forming a Data Governance Council – Representatives from both academic and administrative areas should be a part of the council. The Council’s responsibilities include, but not limited to, creating a data governance charter which establishes policies, rules, and roles related to data. Data resource planning develops a blueprint for data and relationship between those data across the organization. Macro-level data planning done via enterprise data modeling approach depicts the organization and its data requirements. It identifies useful data and valuable information, where they come from, how are they used and how much is expected (Brown et al. 2011, 95-113).
b) Identify Data Stewards – They are responsible for the quality of data within their own realm. Their roles include identifying all the values, naming them uniquely and defining them, set and check quality standards for each type of data and granting access and clearance to use those data. Regular audits must also be done in order to ensure integrity.
c) Security around Data Sources – Data stewards defines users and roles to access data. They also institute a process through which data access request can be reviewed and addressed quickly. Data and application software must be managed as separate entities. This will help keep data secure and is not affected by application changes. The ISO/IEC 27002 information security standard (formerly ISO 17799), which covers security issues relevant to data management, will be followed by us.
2) Data Accessibility – We need a simple yet effective way of sharing data across various departments. Database design should assist in data retrieval and data reporting in an efficient manner that can be easily understood by stakeholders.
Data Warehouse is a method to make data accessible to many people. Data is cleaned and organized using the master data management approach (mainly with the integration approach) and then stored in the database (Brown et al. 2011, 95-113). Functional representatives from each department are responsible for data review and report requests.
Data analysis and data presentation provide data and information to authorized persons. The applications for this can draw upon any data from the database. It can be a summary, comparison of past reports, graphs etc.
3) Data Knowledge – Data stewards offer functional and technical training in their departments. They also share information about policies, procedures, and resources. The naming and definitions for each data type must be compiled and shared across all divisions and campuses. This is a called metadata, which unambiguously describes data for our university. Creating and maintaining high-quality metadata is necessary to ensure quality data. All the data should be stored in the metadata repository or data dictionary/directory. This metadata repository is also used to access and authorize the use of data (Brown et al. 2011, 95-113).
Periodic copies of the database are made to overcome any mishap. If a database is damaged due to software or hardware issues, it can be restored using the backup. Data governance council must also decide, on legal and other grounds, how long the data should be archived. E.g. Transcripts of students will have to be preserved forever but financial records are legally stored for 7 years. According to the type of data, disposal policies will be made.
Data Analysis and Reporting
We need data analytics on a wide range of administrative and operational data to assess institutional performance and progress in order to predict future performance and identify potential issues (Framework adapted from Daniel, 2014). This is described in four ways-
· Institutional analytics
Institutional investigation refers to an assortment of operational information that is broken down to help with successful choices about making upgrades at the institutional level. It includes assessment policy analytics, instructional analytics, and structural analytics. E.g. – infrastructure improvements like upgrading labs, facilities, etc.
· Information technology (IT) analytics
IT analytics aims at integrating data from a various data sources, such as student information, management, and alumni records, etc. It helps in developing data modeling and reveals obstacles regarding accessibility and usability. It can also track the effectiveness of solutions implemented to overcome any obstacles regarding data management. E.g. – upgrading network access for online portals, etc.
· Academic analytics
Academic analytics provides an overview of the educational curriculum. It also helps in formulating answers to improve current academia. Standardized analytics tools provide data to administrators that can aid in the decision-making process and provide a benchmark to compare with other educational organizations. E.g. – adding or removing course offerings by the university depending on current technology trends, areas to concentrate on research work, etc.
· Learning analytics
Learning analysis focuses on the learning process itself. It is largely concerned with improving learner success by analyzing the relationship between the learner, the educator, the study material and any other persuading external factors. This analysis helps in improving courses and learning processes. E.g. – revising the current coursework, updating reading materials, etc.
The cost to build an Enterprise Resource Planning can be between $10,000 – $150,000 (Hutchison, 2015). Other costs to consider include:
I. Cost of Database Management – It is between $15,000 – $65,000
II. Infrastructure Costs – They run between 10% and 20% of the implementation cost
III. Software costs – range from 15% to 30% of the implementation cost
IV. Cost of Human Resources – more than 50% of the total implementation budget
V. Recurrent Costs – The standard license renewal fees is 10% to 15% of the software costs
Implementation cost will be high but in the long run it will reduce expenses related to data management. With advancement in technologies, data storage costs keep decreasing, but unmanaged data volume keeps increasing exponentially every year due to smart devices and upgrades to current infrastructure. So, implementing a data management policy now will help us adapt through growing technologies.
1. Schneider, Christie 2016. The biggest data challenges that you might not even know you have. IBM. May 25, 2016. Accessed Jan 19, 2018.
2. Import.io, 2017. All the Best Big Data Tools and How to Use Them – Big Data. Import.io May 4, 2017. Accessed Jan 19, 2018. https://www.import.io/post/best-big-data-tools-use/
3. Flory, Scott and Walker, Aaron 2015. Data Management Recommendations California State University – Channel Islands. Idata Inc, Alexandria VA. December 3, 2015. Accessed Jan 18, 2018.
4. Brown, Carol V., Daniel W. DeHayes, Jeffrey A. Hoffer, E. Wainright Martin, and William C. Perkins. 2011. Managing Information Technology. 7th ed. Upper Saddle River, NJ. Pearson Prentice Hall. Accessed Jan 16, 2018.
5. Daniel, Ben 2014 Big Data and analytics in higher education: Opportunities and challenges, British Journal of Educational Technology, Accessed Jan 20, 2018.
6. Hutchison, Chandler 2015. How Much Does an ERP Implementation Cost? Clients First. Jan 05, 2015. Accessed Jan 20, 2018. http://blog.clientsfirst-ax.com/blog-1/how-much-is-erp