This Themesis Content Page is designed to give you resources for your research and writing project.
The projects that we address here are Portfolio elements. Each Portfolio element demonstrates your ability to work the highest standards possible, and to conceive of a project, carry it through, and present it professionally. This kind of work typically takes about one quarter (ten-twelve weeks) to complete.
This resource collection is divided into three parts:
- Project organization and planning,
- Specific elements of your paper, and
- Making smart choices about “support resources.”
Typical Paper Organization
These are the typical main sections for your paper:
- Title Page (NEVER numbered; page numbers begin on first page with text following this, pages with Table of Contents, etc., are numbered with small Roman letters, e.g., “i,” etc.)
- Abstract
- Introduction
- Literature Review (or Background)
- Data (including data gathering, preprocessing, feature extraction, etc.)
- Methods
- Results
- Discussion
- Conclusions
- Directions for Future Work
- References
- Appendices (as needed)
In addition, you will likely have these smaller sections at the end of your paper; these are BEFORE your Appendices and References and AFTER your Directions for Future Work:
- Funding (typically needed only for a journal or conference publication)
- Acknowledgements (anyone who contributes to your paper but is not an author, e.g. someone with whom you discuss your paper or who gives you a review; include ChatGPT if you worked with it)
- Conflict of Interest Statement (typically needed only for journal publications)
- Data Availability (either indicate how to access the publicly-available data source that you used, or if it is corporate-proprietary, indicate such so that people know that they can NOT get it, or if you created it yourself and are making it available, the GitHub repository)
- Code Availability (same as above; be careful to clearly delineate which code YOU created or what YOUR part was in constructing an architecture, vs. code provided by others – always attribute!)
If you are planning to submit your work to a journal or a conference, check their “Instructions to Authors” FIRST; they will typically have their own preferred format which might be a little different from that presented here.
Project Plan, Week-by-Week
I and many others find it useful to do our “life planning” in terms of quarters instead of years. This can be useful in planning both corporate (or team) goals as well as personal ones.
The book 12-Week Year, by Brian Moran, advocates this kind of planning cycle.
This is an approximate project plan and schedule, which will take (typically) ten weeks to complete.
Plan Your Review Cycle
Sometime during Week 1 and Week 2, you want to build in your project/paper review into your work plan. We’ll offer some review options later. Just plan to get early drafts done sufficiently in advance so that you can show your work to someone else and get feedback.
Plan Your ChatGPT Involvement
Also, plan ahead how much – to what extent – you want to rely on an AI resource such as ChatGPT. Using an AI assistant is acceptable – you have to acknowledge this use, but it’s ok. Just be very, VERY careful. LLMs and tools built on them (such as ChatGPT) STILL HALUCINATE (as of this writing, Jan. 2024), and you need to plan to fact-check them carefully.
Good places to use ChatGPT are:
- Lit Review – ChatGPT can help summarize your references. Be VERY careful to check that any references that ChatGPT (or any LLM-based tool) offers ACTUALLY EXIST. DOUBLE-CHECK the exact reference listing (the citation). Also, you MUST go and read at least portions (Abstract, Intro, glance at the rest) of EACH of your references. (Otherwise, you’ll get caught out someday, and it will be very embarrassing.)
- Discussion – AFTER you’ve obtained your results, and you’re trying to put your work in context, it may be useful to have some conversations with ChatGPT, and use ChatGPT excerpts in your work. Again, be careful. Don’t include anything that you, yourself, would not say. Be ready to back up your points.
Where I DO NOT favor using ChatGPT (or any other tool) is in your Introduction and Problem Statement. You are seeking to present your own unique, definitive, and distinguishing work. You cannot afford to have your voice weakened or your clarity diminished by relying too much on what others say.
It’s better to go for a long walk (or a series of walks), come back and noodle things, and do multiple iterations than to get an AI involved in this.
AFTER you have your Intro & Problem Statement fairly well worked out, you can consider what ChatGPT has to say.
Again, be careful. You want to be uniquely you – not “writing by committee.”
Project Plan Overview
- Week 1: Identify your project, collect resources, draft paper outline. (Optional: Draft YouTube or audio-vis presentation outline.) (Start to) get data.
- Week 2: Clarify project goals, obtain data. (Start to) develop your Literature Review. Craft first draft of your Introduction.
- Weeks 3 – 10: (To be continued)
Week 1: Identify Your Project, Collect Resources, Draft Paper Outline, (Start to) Get Data
Devise Initial Problem Statement
You might start with a general notion of what it is that you intend to investigate, and clarify your exact focus (your Problem Statement) as you go along. Still, do your best to devise an early Problem Statement draft.
Your Problem Statement should show up in your:
- Abstract,
- Introduction, and
- Discussion.
Craft Your Title Page
Your title page is something that you can do IMMEDIATELY.
This YouTube illustrates crafting your title page.
Put In Some References
Building your reference list takes time. Get the first few – the ones that you KNOW that you want – in there early.
In Northwestern’s MSDS program, we use the Chicago Manual of Style (Author/Date) style for the paper (overall) and for references.
Get Data
Getting your data, or creating it, and doing the data cleanse/preprocessing/feature extraction work will consume MOST of these next few weeks.
Get started early. Do this in parallel with writing, so that the early portion of your paper (Intro and Lit Rev) are pretty much DONE by the time that you start putting your data into the code and writing your Data section.
End-of-Week-1 Objective
Paper: By the end of Week 1, it would be ideal to have your paper outlined in the style and form that you will use for the final version. Have section headers with stub paragraphs beneath them.
Data: Also, the earlier you can obtain and/or create your data, the better. It ALWAYS takes more time than you think to do data cleansing and preparations.
References / Prior Work: Find ALL available material that you have on hand, such as references that you KNOW that you’ll include, and get them in.
If you are building on your own prior work, be certain to cite yourself.
Also, if you’re building on prior work and have certain figures that you will use, include them now. Revise as needed for your current work.
Optional: IN EACH SECTION, put down your ideal/planned date for getting that section draft completed. Use this for your project planning.
Week 2: Clarify Project Goals, Obtain Data. (Start to) Develop your Literature Review. Craft (first draft of your) Introduction.
- Construct a first pass at your Problem Statement. This is the lead paragraph in your Introduction.
- Construct a first pass at your Abstract.
Your Abstract and Introduction should BOTH lead with a strong, clear Problem Statement. This is the most important task that you’ll have in writing your paper.
You must make a clear and compelling case immediately, because people have so many choices. This means:
- The first sentence in your Abstract should be your Problem Statement, with a clear indication that this is important. See minute 4 ++ in the YouTube below.
- The first sentence and first paragraph in your Introduction should present and clarify your Problem Statement.
- It can take multiple revisions to get clarity; plan for at least two major revision-rounds.
This YouTube shows how I revised my Abstract while developing a project.