Back to Homepage
Credit: Public Domain
Milestone 6: Archive
Benjamin Xie & Gregory L. Nelson
For your final project, you will create 2 things:
- an artifact that decision-makers (your intended audience) could use to inform a decision you define.
- a technical description which other data scientists could use to understand and build off of your analysis
Deliverable 1: Artifact
The target audience for the artifact will be the decision makers in your decision context.
They should be able to interpret and engage with your artifact to help them make more informed decisions.
The challenge to making an effective artifact, much like making an effective visual, is about ensuring expressiveness and effectiveness by balancing what detail to show.
Your target audience should know just enough to be more confident they are making a decision that better optimizes what they want.
They may be confused or misled by excessive detail or make incorrect conclusions with insufficient detail.
Finding the appropriate balance is dependent on your decision context and target audience.
Deliverable 2: Technical Description
The target audience for the technical description is other data scientists who wish to interpret, critique, and build off of your work.
It will be a written report in the form of a page on your GitHub wiki and should provide sufficient detail so somebody can understand your analysis and interpret it.
This assignment is out of 16 points with options for extra credit. You should be able to access it from your team's GitHub wiki:
From the perspective of a decision-maker (your intended audience), we will your artifact as follows:
- 1 point: I understand the decision this artifact is supposed to inform.
- 0.5 points: I understand the method for supporting the decision and why it is appropriate.
- 0.5 points: I understand limitations for the method.
- 2 points: I understand how to use the results/method to inform the decision.
- 1 point: The design and presentation has a high quality feel (formatting, aesthetics, information design).
- 1 point: Visualizations are expressive and effective to intended audience.
- 1 point: I understand the impact of different choices on potential outcomes for the decision.
- Extra credit (up to 2 points): This artifact is complete and polished enough to inform real decision-makers.
From the perspective of another data scientist, we will grade your technical description as follows:
- 1 point: I understand who the target audience and decision context is for this artifact.
- 1 point: Dataset and variables (predictive, outcome) are described.
- 0.5 points: Data cleaning process is described.
- 0.5 point: Team justifies use of method for supporting the decision.
- 4 point: Analysis is high quality and thorough.
- 0.5 point: Assumptions and limitations clearly discussed.
- 1 point: Project is well-documented (comments in code, wiki).
- 0.5 points: Potential future work is described.
- Extra credit (up to 1 point): Team used a method not taught in class to improve their analysis and justified their choice to use this method