You and your team will pitch and execute a text mining project on the data of your choice to derive insights valuable for a business, a public policy-maker, or a charity. You will also review a project pitch created by another team in your class.

PROJECT TEAMS

You will work on the project as part of a student team, as described in the Syllabus.

DATA

You need to obtain text data for your project. You are welcome to use one of the available datasets (see links below). You can work with any dataset as long as your access to and use of the dataset are

not in violation of its terms of use.

Some sources of text data to consider:

•   Datasets on Kaggle and GitHub.

•   Yelp Reviews (available here:http://bit.ly/2TBI9m2);

OBJECTIVES AND METHODS

You will set objectives and choose methods of analysis for your project.

Project objectives should reflect insights you plan to get and business outcomes your project helps achieve. In other words, you need to explain what information you plan to obtain and how that information affects business outcomes. The insights should be actionable andfeasible given the nature of your data and nature of the business.

o Project objectives should not be technical, such as “we plan to do topic modeling using

car reviews.” Topic modeling is just a technical tool you use to derive insights. You must use two (2) text mining methods/techniques for the analysis stage:

o For example, you can choose topic modeling as the first technique and one of the methods of sentiment analysis  as the  second  technique. You  can  also  choose  to  compare two techniques of sentiment analysis.

▪   Note: methods of data pre-processing do not count towards the required 2 analysis methods.

o If needed, feel free to use techniques you learned in other classes for your analysis.

DELIVERABLES

There are four project-related deliverables:

1.   Project Scope (a pitch in class + a deck of 3 slides)

2.   Peer Feedback (evaluation of another team’s Project Scope)

3.   Final Report

4.   Team Evaluation

Details and tips for the deliverables are below.

Project Scope: Slide Deck

Project Scope is a short description of project objectives, your data and analysis methods. Your ultimate goal for Project Scope is to explain:

1)  The insights (information) that you plan to get

2)  How these insights help improve business outcomes

3)  Why your chosen methods and your data allow you to get the above insights

You will pitch your Project Scope and hold a Q&A session in front of the class during the last lab to get feedback on your project idea.

The learning objective of Project Scope activities is for you to get experience in: 1) delivering short, focused business presentations, and 2) giving and receiving peer feedback.

Rules:

▪   Your Project Scope should:

1)   Outline the objectives for your project:

-     What insights do you plan to get?

-     How can an organization use those insights? How does the information you plan to obtain improve business outcomes?

2)  Briefly describe your data:

-     How and when were the data collected?

-     What are the shortcomings of the data (e.g., possible sample biases), if any?

-     Think about ways to visualize some important aspects of your data.

3)   Explain what methods you will use and why.

4)  Include photos of all Team members on the title slide to introduce your team to the class.

5)  Have an active weblink to your data source.

▪   Do not include the results of your analysis in your Project Scope even if you have them; instead, spend time on clearly presenting your project idea.

▪   Electronic deliverable: a deck of 3 slides, not including the title and thank-you slides.

▪   Submit on Canvas by the deadline (see Important Dates below).

Tips:

o Create a descriptive project title to convey information about what you do in your project:

o For example, “Ford car reviews” is not a good title.

o “Extracting brand perception from consumer reviews of Ford cars to improve product quality” is better (it might lack spice, but you are free to have a creative title as long as  it is descriptive ☺)

o Consider preparing a “McKinsey-style” slide deck, a concise but information-dense set of slides which combine visuals and text.

o On Canvas, you can find a student-generated deck of slides with helpful tips on preparing a “McKinsey-style” presentation.

o Your work on the project should be well under way at the time of your pitch:

o This way, you can make sure your ideas for the project are well developed.

Project Scope: In-Class Pitching

You and your team will pitch your Project Scope in front of the class during the last lab session.

Based on your slides and the pitch, you will receive written comments and suggestions from 3 to 5 of your classmates (see Peer Feedback below). You should incorporate the suggestions from your peers if they are helpful since it will improve your project.

Rules:

▪   You will use Project Scope slides for your in-class project pitch (you will not be able to update the slides after you submit them on Canvas).

▪   Your team will have about 5 minutes for your pitch: exact timing will be confirmed later in class.

▪   Everyone on your team must participate in delivering the pitch.

Peer Feedback

You will provide constructive comments and suggestions for one (1) Project Scope created by another team in your section of the course.

This is an individual, not team-based, assignment: your teammates will be giving their feedback to a team different from the one you are assigned to review.

Rules and important details:

▪   Ask questions about the team’s project during their Q&A session: it will allow you to provide a quality feedback to the team.

▪   Deliverable: 1-page write-up (Times New Roman, size 11, single-spaced).

▪   How to submit: You will receive an email invitation from Canvas to submit your Peer Feedback.

▪   In your Peer Feedback you should:

-     Write for and address your comments to the team: they will receive your review.

-     Describe what you like about the team’s project.

-     Comment on the issues that the project might have. Think through the following:

o Are the project objectives managerially relevant?

o How realistic are the project objectives given the data?

o Any selection issues with the data that are relevant for the intended analysis?

o Are the selected methods able to deliver on project objectives?

-     Brainstorm: How would you improve the project?

Tips:

o You will have access to the assigned team’s Project Scope before the pitch: do get familiar with it beforehand to know which team is yours to review and think through questions you might    want to ask.

o Go to the source of the team’s data to get more familiar with the dataset.

o Be generous with your advice and constructive in your criticism.

Final Report

The final report for your project will describe your data and objectives in more detail and include   rationale for the project design, explanation of your methodology, project results and discussion of business decisions the organization can make based on the project results.

Rules:

▪   Final Report must be 2.5-3 pages (not including the appendices with tables and plots).

▪   The report should include the following sections:

1)   Executive summary (one paragraph summarizing your project)

2)  Project objectives

3)  Data description

4)  Methodology:

o Describe your methodology and discuss the pros and cons for using this methodology to achieve your project objectives.

5)  Results and their discussion

6)   Conclusion:

o Provide recommendations for the parties interested in your results;

o Indicate possible shortcomings of your work and ways to overcome them.

7)  Appendix: tables and plots

8)  References: make sure you cite properly the sources you used, if any (any citation style is fine).

▪   Please do not write your names or student ID numbers on your report.

▪   Writing style:

o Use bullet points only where appropriate.

o Be brief and to the point.

▪   Formatting and technical details:

o The report must be in the .pdf format.

o Font and spacing: Times New Roman font, font size 11, single-spaced.

o Do not include a separate title page.

o Plots and tables must have titles and enumerated, be referenced in the text and placed into an appendix.

o Provide a link to your data source.

o Your Python script, clean and commented, must be a Jupyter file.

o Upload your report file and your Jupyter file to Canvas as 2 separate files (do not use archives, such as zip).

GRADING

The project and related activities weigh 40% of your grade for this course.

Project Scope (pitch + deck of slides)  10%

Peer Feedback  10%

Final Report   20%