TrackMaven, a marketing analytics group, has asked for your help with Condé Nast, one of their clients.
Problem: Identify topics and image attributes that most and least drive engagement for your assigned channel(s). Note that you may want to define one or more metrics that measure "engagement."
Data: The data can be found here. (Reading this data into a Pandas dataframe took approximately 3 minutes. You should have 758,763 rows and 11 columns.)
Deliverables: 1) A presentation to TrackMaven stakeholders. 2) Cleaned code for delivery to TrackMaven.
Deadline: December 8, 2017.
Questions: Questions should be sent via Slack to @matt.brems. Questions should be specific, brief, and formatted so that they can be directly sent to our contacts at TrackMaven.
| Team # | Team Members | Assigned Channel(s) |
|---|---|---|
| Team 1 | Dale, Kyle, Natalie, Diego, Matt | |
| Team 2 | Chris, Joel, Scott, Anirudh, Alex | Pinterest & Blog Posts |
| Team 3 | Ben, Chukwudi, Steve, James | Facebook & Instagram |
You will be able to subset the provided data so that you only focus on your assigned channels, although you may find some value in examining data in other channels.
Data science is a field in which we apply data to solve real-world problems. Therefore, projects and presentations are means by which we can assess your ability to solve real-world problems in a data-driven manner.
When evaluating projects, there are four areas on which your instructors focus.
-
Project Requirements: Did you meet all project requirements? In answering this question, your instructors want to assess how well you met the project requirements as established. These will generally be laid out in the project readme.
-
Audience: Is your presentation appropriate for the stakeholder? In answering this question, your instructors want to assess how well you present your results to stakeholders. For example:
- Did you frame the problem appropriately for the audience?
- Did you use the appropriate level of technical language for your audience?
- Did you effectively use your time, or did you encounter an issue such as going significantly beyond or under the allotted time or rushing to conclude the presentation in the allotted time?
- Did you present effectively, or were there things that detract from the overall presentation such as not speaking loudly enough for the audience or repeating oneself?
- Methods: Are your methods appropriate for solving the problem? In answering this question, your instructors want to assess how well you have applied data science methodology to the problem at hand. For example:
- Did you make well-reasoned modeling choices, or is there clear evidence that the model is inadequate or improper?
- Are you able to clearly defend your methodological decisions and results?
- Did you generalize your results properly, or were your conclusions/inferences improper or fallacious?
- Value: Have you provided value to the stakeholder through clear, data-driven recommendations? In answering this question, your instructors want to assess the value you provide to the stakeholder as a data scientist. For example:
- Did you answer the problem posed to you?
- Did you make your recommendations clear, or were the recommendations unclear?
- Were your recommendations data-driven and based on the results of your work?
You will earn a score for each of the four areas mentioned above.
- Project Requirements: You may earn a score of 0 or 1. You will earn a score of 1 if all project requirements are met. Otherwise, you will earn a score of 0.
- Audience: You may earn a score between 0 and 3. A score of 0 indicates that your presentation is inappropriate for the stakeholder. A score of 1 indicates that at least part of your presentation should be non-trivially reworked to be more appropriate for the stakeholder. A score of 2 indicates that there are few to no areas of your presentation that should be reworked. A score of 3 indicates that your presentation is consistently appropriate for the stakeholder and serves as a model for future presentations.
- Methods: You may earn a score between 0 and 3. A score of 0 indicates that your methods are inappropriate. A score of 1 indicates that your methods are somewhat inappropriate, that justification for methodological decisions is lacking, and/or that your conclusions do not follow from the methods. A score of 2 indicates that your methods are appropriate, justification is sufficient/strong, and your conclusions follow well from the methods. A score of 3 indicates that your methods are excellent, strongly defended, and serves model for future presentations.
- Value: You may earn a score between 0 and 3. A score of 0 indicates that you provide little to no value to the stakeholder. A score of 1 indicates that the value you provide to the stakeholder is substantially less than expected by not answering the problem, not providing clear recommendations to the stakeholder, and/or providing recommendations that were not data-driven. A score of 2 indicates that the value you provide to the stakeholder is on par with the expectation of providing clear, data-driven recommendations that directly answer the problem posed. A score of 3 indicates that the value you provide to the stakeholder is beyond what is expected and serves as a model for future presentations.
Your final grade will be calculated as follows:
- If any project requirement is not met, the final grade is 'Fail' with a score of 0.
- If all project requirements are met, then the final grade is 'Pass' with a score calculated by summing the above scores. Therefore, if all project requirements are met, the final score will be between a 1 and 10.