Jacob J. Walker's Blog

Scholarly Thoughts, Research, and Journalism for Informal Peer Review

My Response to a Job Application Question about Developing Reports from Databases

without comments

While my major focus right now is on getting up and running the Highlands Community Charter and Technical Schools, as that is the most likely place that I will be able to obtain employment, I am also applying to a few other jobs that are within the scope of my “usual” employment.   One of these is for the position of a programmer/analyst with CSIS.  (Yes, the same CSIS that I’ve been talking about in my past posts.)  I generally try to answer these types of questions in an educational manner, such that it not only talks about the experience that I have, but also clearly demonstrates my knowledge.   Because of this, I thought that some of these answers are worthwhile sharing here, as they can help others to understand some of the skills and processes used in real-world computer science.  So I will be writing a series of blog articles that contain my answers.

The first question I will address is:

Describe your experience in the creation and maintenance of online or batch report development. Be specific as to the technologies used, the role(s) that you performed, and your deliverables.

To do report development and maintenance well, it requires competence in several disparate skill sets.  First there is the need for the skills of a consultant to understand the goals of the report; otherwise it is quite easy to develop a report that does not return the data the end-user wishes to have.

As Department Chairperson of Career Technical Education with Twin Rivers Adult School, and having been in a member of several standing committees, such as one about student retention, I have seen how data is used to make data driven decisions, and understand the purpose for creating reports for others, and often acted in the role as an internal consultant.

Next, there is a need for skills and knowledge of data science, with an understanding of the underlying data that will be part of the report, and the limitations that this data may have, and assumptions that may be made in the process.   There is also a need to recognize GIGO (Garbage In, Garbage Out) problems that often exist with database systems.

Having been in the role as the primary institutional researcher for Twin Rivers Adult School, I had to gain a good understanding of our data, the limitations, and what assumptions could be made in the process that would produce sufficiently accurate results, and also to be wary of where GIGO problems might exist that would cause inaccurate results, despite having queries or reports being properly created.  Further, through my own self-study, having gone through a Master’s program in Education, and now being in a doctoral program, I have done extensive work with research methodologies.

Next, there is a need for the technical skills to extract, transform, and display the data.  I have been programming both personally and professionally (with small scale projects) for over 20 years.  I have sufficient proficiency in a variety of programming languages and data technologies, including T-SQL, Python, Java, PHP, and VBA.  Beyond my work with programming languages, I have had extensive experience using Microsoft Excel to connect to an SQL database, retrieve data into an Excel table via an SQL query, and then to analyze that data using pivot tables, and ultimately present the data using appropriate charts and worksheets that were designed to be in a report format. For example, one of the deliverables of a report from Excel was generating a student award letter, which based upon the students qualification data from the other databases and the schedule data that was in the spreadsheet, that a report would generate showing how much the student would receive in Pell Grants, and how often they would be paid their “cost of living” if they had a Pell Grant greater than the cost of their program.

Finally, there is a need for skills and knowledge of making the data easy to understand for the end user. This involves both good visual design and communication skills to make sure that the data is presented in a way that can help the end user understand what is being shown in an accurate manner.  For example, as a deliverable for my administrators, I have often created line graphs to show correlations between variables.  Using the feature where Excel can show two Y-axes with the same X-axis, two variables can be plotted, and if they clearly rise and fall together, then there is a clear correlation, and can be easier to understand than trying to fit an equation to a scatterplot.   This technique worked very well to demonstrate that one of the primary variables that affected total average daily attendance was simply the number of courses offered, for while there was a variability of attendance within each course, this did not have as much influence on the total attendance as simply having the courses.

Another deliverable was an investigation that I did for the committee that looked at student retention.  In the adult school, we had many open entry courses, where new students could start every week.  Analyzing the data, and presenting a graph of various programs, it was clear that students who started in pairs would do better than those who started alone, and better than those who started in 3’s.  Although, this may have been due to couples (husband and wife, or friends) starting together, and this could not easily be sorted out from the data.  But it did lend credence to potentially implementing a strategy of having students buddy up in a course, to help the students improve their persistence in the courses.

I also was the primary accreditation liaison for the adult school, and as a deliverable I would write much of the accreditation report based upon the research I did with interviewing faculty, leading meetings, and querying our databases.  Often data would be presented in pie charts in these reports, and I would be sure to follow the best practice of using a 2d chart, as a 3d chart would generally cause a visual skewing of the data.

Post Revisions:

Written by Jacob Walker

July 10th, 2013 at 7:05 am

Leave a Reply

%d bloggers like this: