Milestone 7: Project Fair

Gregory L. Nelson (some material made with Benjamin Xie)

The purpose of this milestone is to gain practice with communicating your data science briefly in an "elevator pitch" style. These skills will help you describe your work during interviews, networking, on the job to your boss and especially to the leaders of your company; doing this well allows people to remember what you did, why it was important, and remember you.

This involves story-telling and describing and interpreting your data science analysis. Interpreting your analysis is translating technical details into what they mean for what decision-makers care about e.g. outcomes. Interpreting your analysis means not just telling people what coefficients your model had or giving lots of details about the technical parts of your analysis they may not understand.

For this assignment, you will prepare an elevator pitch, with two different versions. This will give you practice in adapting your communication to different audiences - someone unfamiliar with your domain, and someone familiar with it.

Following your pitch, you should be prepared to answer some follow-up questions. This will give you some practice at thinking and communicating effective answers with minimal preparation time.

Format

Your group brings: laptop, your artifact and technical description, anything else you think you need prepared. Visitors will go around to projects, introducing themselves. You will ask them about their background knowledge of your domain. Given their background, you will then give your "elevator pitch" 1) in 2 minutes if they are familiar with your domain, and 2) 3 minutes if they are unfamiliar. Afterwards, they may ask some clarifying questions, then give you feedback. Your grade will be determined by Greg coming to your group to listen to each pitch closer to the end of the project fair.

See the Grading Criteria to see a more detailed breakdown of how we will assess you.

Common Missteps

Often times, great people struggle to communicate well. Below are a few common mistakes to communicating data science. We hope we don't see them at the project fair!

Not stating the decision context and different choices.
Presenting data or analysis in graphs/charts without interpreting it (you should describe it like you interpreted in tab 1.2 in PM spreadhseet).
Not connecting your analysis to the decision context and question to answer.
Running out of time without finishing all the points.
Having people look at some visual that is too complicated, and your audience gets too fixated on reading/exploring it vs. listening to you. (split attention effect)

The best way to avoid these mistakes is to 1) focus on the big picture objective of what you want to convey, 2) practice delivering your pitches, 3) cut out everything that isn't required in conveying the big picture. See Further reading for resources on speaking (in the contxt of presentations but the principles generalize to pitches).

Grading Criteria

This assignment is out of 8 points (with an opportunity for extra credit) and you will be graded on your elevator pitches and how you answer follow-up questions related to each point. If you create any slides or visuals to display (besides project artifact and technical description), include them (or a link to them) in your project GitHub or project wiki.

For the pitch to someone unfamiliar with your domain: (4 points, equally weighted)

2/3 point: Decision Context. You convey the decision context surrounding your study and why it is important.
2/3 point: Data. You convey where your data came from and how it was used.
2/3 point: Method. You briefly explain your analysis methods and choices.
2/3 point: Results. You describe and interpret the results of your analysis.
2/3 point: Informing a decision. You explain how your target audience may use these results to inform a decision.
2/3 point: Limitations. You describe any significant assumptions or shortcomings of your analysis.

For the pitch to someone familiar with your domain, cut back on the decision context and explaining what different data field / domain concepts are - you can just use them assuming your audience knows what they mean. The grading criteria is basically the same as above: (4 points)

0.33 point: Decision Context. You convey the decision context surrounding your study and why it is important.
0.33 point: Data. You convey where your data came from and how it was used.
0.33 point: Method. You briefly explain your analysis methods and choices.
1 point: Results. You describe and interpret the results of your analysis.
1 point: Informing a decision. You explain how your target audience may use these results to inform a decision.
1 points: Limitations. You describe any significant assumptions or shortcomings of your analysis.

Extra credit (0-1 point): Give another version of your elevator pitch to a person familiar with your domain, only using 30-45 seconds.