2

I am writing my Master thesis report. My thesis was research and development of a test data generator for a particular domain in an organization. The developed application will generate test data in XML and JSON format according to user requirements and design. These data can be very large.

For the demonstration of my results in Thesis report, I have designed a simple model and generating the test data for that design but still, the test data generated is pretty large so I have included them in the Appendix chapter of the result. It's taking around 20-30 pages, I have modelled the simplest design so there is no way for me to reduce the generated test data by simplifying the model further.

My professor has told me that this is not a good way of displaying the results. So I wanted to know what are the better ways of showing/visualizing your results/test data in the Master thesis report?

As I am writing the report in LaTeX format I have included them as listings in the Appendix section. Please provide me with some solution.

3
  • Do you have access to previous dissertations in your field (ideally supervised by the same professor)? You could ask your supervisor or at the university library.
    – astronat
    Oct 22, 2020 at 10:22
  • Thanks for the response. Actually, I have one report from my senior who was supervised by the same professor but he only has some database query and table structure. He has not presented any output result screenshot within the Appendix section. Since my topic is very different and the result is completely different I have included the output in the Appendix section. Oct 22, 2020 at 10:40
  • I don't think you have a LaTeX issue here, but a general issue if organizing your writing and thoughts.
    – henning
    Oct 22, 2020 at 19:49

1 Answer 1

2

The full data is the main outcome for the organization, but it's not an important part for your report. Of course you can show examples/samples, but what matters in your report is to explain clearly:

  • The motivations: why is this data generator needed? What can be done with it that couldn't be done before? This probably requires explainining the context of the work: what this organization does, what is this domain, why does it matter, etc.
  • The state of the art: which existing methods are used for this kind of problems in general? What are their advantages and disadvantages?
  • The method/process: how does it work? Provide all the details necessary the system. If you made particular implementation choices, explain why this choices and not another option. Is it related to existing methods? If yes compare to these methods.
  • The outcome: analyze the results in detail with statistics, graphs or whatever means helps the reader understand the characteristics of the data. It should be a synthetic overview, not a dump of the full dataset. Prove that the obtained data satisfies the requirements, for instance by comparing to other datasets. If relevant show the impact of your implementation choices through some examples.

The point of the report is to explain your work to somebody who is not an expert in this particular problem. Presenting the reasoning that you followed is much more important and interesting than the data itself.

3
  • I think this is the right answer. Perhaps the raw data can be made available in some repository maintained by the university, should anyone need to look at it. Oct 22, 2020 at 16:33
  • Thanks a lot for the response. I completely understand your answer and agree with it. However, the main goal of the research was to generate the test data in XML and JSON format so I am not exactly sure how can I visualize that in the form of graph or chart. If possible can you please provide me with some suggestions or example on achieving this. Oct 23, 2020 at 7:37
  • XML/JSON is just a technical format. You should explain why this format is used (e.g. a standard compatible with many software tools, etc.), but I assume that the main point is about what this data represents. For instance if the data represents customers you could demonstrate that the data correctly represents the distribution of male/female among customers, or the average amount a customer spends, etc. This kind of characteristics can be represented with graphs, tables, etc.
    – Erwan
    Oct 23, 2020 at 10:20

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .