MGMT 690 MS BAIM Industry Practicum Spring 2022

Task List #4

Due: Sunday, February 13th at 11:59pm EST

All teams appear to be working well. The INFORMS Business Analytics conference has changed the
poster deadline date to March 4 th to upload a poster abstract and poster draft, and Mary 11th to upload
a finalized draft if accepted. Due to the extra couple weeks, we got from this deadline shift I won’t ask
you to upload a poster abstract for this task list.

1. [1 point] Create a folder called Archive in your team’s Box folder. In this folder create a backup
copy of your current slides and paper.

2. [10 points] In your working paper, address any comments posted from your instructor(s) so far
in your paper. Most comments will come from the Introduction section, Literature Review, or
Data sections of the paper.

3. [1 point] On April 8 TH or 9th you will present your project at the Virtual Midwest Decision
Sciences Institute (MWDSI) conference (conference link) where you will present a 15-minute
PowerPoint presentation. We would like all team members to present a part of the overall
presentation. However, we will only register and pay for one member of the team to be the
“primary presenter/attendee”. Unlike the INFORMS Business Analytics conference, this is a
required event, and will not cost anybody on your team any money. You just need to identify
who will be “primary presenter” for your team in column H of this Google sheet link below.
Once this task list is due we will register this person from your team. They will submit your
paper at a future date.

https://docs.google.com/spreadsheets/d/1030oerGoPcnXdSyiRrSxN01xwET0dd-
oRVu4qZwWzyQ/edit?usp=sharing

If your paper is accepted and presented at the conference, all your teammates will have a
‘conference proceeding’ publication you can to your resume or LinkedIn profile under
‘Publications’. Here are a couple examples from former BAIMers:

https://www.linkedin.com/in/mohinder-goyal/
MGMT 690 MS BAIM Industry Practicum Spring 2022

https://www.linkedin.com/in/xuanmainguyen/

Most professional master students don’t write and submit papers at conferences, but MS
BAIMers do! This will help provide additional evidence of the technical rigor you bring to the
table. Some students in the past have also put their work under the ‘Projects’ and ‘Experience’
sections of the profile. See these examples:
 Jiangtao Xie, currently works at Tredence (https://www.linkedin.com/in/jiangtao-xie-
lotto/)
 Sachin Arakeri, currently works at Amazon
(https://www.linkedin.com/in/sachinarakeri/)
 Stefanie Walsh, currently works at Altus Group (https://www.linkedin.com/in/stefanie-
walsh/)

If and when you add IP project stuff to your LinkedIn profile or resume it is important you do
NOT violate the NDA you signed. You will notice most previous students say “Undisclosed
Industry Partner” or similar verbiage, “Undisclosed National Retailer” etc. When people see that
they know you signed an NDA. There is nothing fishy about it. You can only indicate the
partner’s name you worked with IF and ONLY IF they send you written confirmation (email is
fine) that they are okay with it. You should wait to ask to see if this is a possibility once you have
finished the project. Only at that time can the client make an informed decision and or discuss it
with their legal team.

4. [15 points] Finalize the literature review section of your paper. The expectation here is you have
identified and summarized approximately 5-10 papers that are closely related to the goals of
your project. You will have at least one table that provides the reader a summary of what you
wrote about those papers, but the writing will not be IN the table. Your table(s) might look
something like this:

[Example 1: paper link]

[Example 2: paper link]

[Example 3: paper link]

-
MGMT 690 MS BAIM Industry Practicum Spring 2022

5. [18 points] Divide and conquer! Discuss with your team and assign each person a data analysis,
modeling, or set of experiments to run that have a clear focus on achieving your research
questions/objectives. If you’re working on a project where many prediction models need to be
tried to identify drivers, best prediction, etc. make sure you’re not doing the same experiments.
Things to think about, discuss, and document before doing any experiments:
 Should we be using the same train/test data sets? How is best to generate those?
 What seed (e.g., 1234) should we all be using? Repeatability is key in the work that you
will be doing.
 What statistical metrics should we all be capturing (e.g., does training time matter?,
does scoring time matter?). Discuss the stats that you plan to capture with the client and
try to justify why you’re going to record those in your experiments. They might suggest
you capture others – then do it. Those stats might be most important to them, but you
can record anything you want. From my experience, the more logging/benchmarking of
various stats the better. It’s easier to delete results you don’t need than try to go back
and capture them later.
 After you train a model, does it need to be saved? For example, if you let a model train
for a day, will you want to score it on various datasets later? Does the client want them?
Consider naming and dumping them out into a Box folder. Remember you have
unlimited space.
 After you train a model do you need to capture the features used for that model, do you
need to capture the beta coefficients, or are there things about your model(s) you might
want to analyze later. Play devil’s advocate. Are there values you might save that might
have potential to come in handy later?
 How can we easily dump and share results across multiple team members? Should we
dump data into a relationship database? Should we dump output files that are in the
same format to a folder? Should we sit a PowerBI or Tableau dashboard on top of that
results DB table or folder that will automatically update?

When you decide to do experiments start with smaller datasets and simpler models and let
what you learn from each run guide you on what to try next. Document stuff in a slide in your
deck so you don’t forget what you did a couple days later.

I want to see in your slide deck how your team addressed the above bullet points (assuming
those are relevant to your project – software teams show us what you considered).

6. [5 points] Create a figure that you can add into the Introduction of your paper and poster later
that aligns with that section. Remember in the introduction you’re motivating your work. Setting
the stage for the audience of why this problem matters and what you are focusing on. In many
academic papers you probably noticed figures were lacking. You want to be a great story teller
and your audience will be your future employer!

Good examples:
 Team created a figure in Excel that showed the idea of a price elasticity curve with
annotations [link]
 Team was make it clear how cost of labor relates to sales and partial profit (concepts
that might be confusing if not familiar with the area) [link]
MGMT 690 MS BAIM Industry Practicum Spring 2022

 Team used images from the project to help show the process [link]
 Team used images that make it a clear to the audience the contrast of the image
matters [link]
Bad examples:
 Team used clip art [link]
 Too wordy; lack of supporting images [link]

Team feedback:
 To think about: I would really like to see the quality of the posters to improve to a more
polished poster-masters level this year. It’s not that the posters students have presented in the
past were poor – they actually won several competitions, but I want a conscience effort to
create visuals that really show you spent time on your work. Let me share some examples I
found online that if your team could do something similar would definitely set you apart at
Analytics or Data Science conferences. Notice the quality of the figures.
o https://solarfuelshub.org/2020-solar-fuels-science-meeting-posters-page
o https://ksqtx.com/news-events/eortc-poster-presentation-comprehensive-
identification-of-novel-therapeutic-targets-for-treatment-of-pd-1-resistant-solid-
tumors-via-a-genome-scale-crispr-cas9-in-vivo-t-cell-screen/
Think about your team presenting your poster. The more effective, readable, and polished
looking visuals you have, the more credible your work will be (even if it’s a bunch of gibberish).
We can put some of those visuals in your paper too!

[5 points] Brief summary about your progress in the following aspects:

 Data cleaning
MGMT 690 MS BAIM Industry Practicum Spring 2022

 Computing issue solved (if you have)?