How will organizations and students share data and analyses?

1. Organizations will post their challenges and accompanying resources on the experiential learning platform Coursolve (http://coursolve.org).  After posting a need on Coursolve (http://coursolve.org/post-need), they can create a thread in the Real-World Projects forum to inform students of their data science need in order to solicit assistance.
2. Students will then have the opportunity to join the Optional Course Project on Coursolve (http://coursolve.org/courseproject/43) by registering and clicking "work on this course project" on the right-hand side.  They can then browse needs (http://coursolve.org/browse-needs) to request to work with organizations whose needs they are interested in helping to address.

Students and organizations will be able to collaborate in the Real-World Projects forum or on Coursolve.  In addition to any solutions or outputs shared on Coursolve, in order to receive credit for the Optional Real-World Assignment, learners will have to upload these submissions to Coursera as well as participate in a peer-review process.  More details on the exact procedure will soon follow.  

What is expected of the organization providing the project?

Here are some tips:

1. Be clear and concise in project descriptions. Projects posted with concise descriptions and a clear articulation of the need + what the ramifications of meeting this need will be are most likely to be selected for analysis.
2. Engage actively in collaborations with students. Responding to emails, forum posts, and other communication in a timely fashion will be fundamental to a successful partnership, as will providing students with the resources they need to complete the projects.
3. Adhere to good data practices. Anonymizing data sets according to existing best practices and following organizational standards for maintaining dataset confidentiality is important. Members of organizations are also encouraged to check in on policies that relate to administering nondisclosure agreements, intellectual property agreements, and other contracts necessary for third-party relationships.
4. Verify student work. As with anything that is produced for your organization, it will be important to verify the analysis and insights students produce against previous research and/or your own experiences and intuitions. Blindly trusting an analysis that ends up being erroneous or misguided may have adverse effects on your organization.

What kinds of help will students be able to offer?

The 8-week course will engage students in topics that span four categories: scalable data manipulation, analytics, communicating results, must-know algorithms and techniques. Students will study a wide range of concepts and technical practices, and have will have the opportunity to apply their knowledge through course projects. The course syllabus offers more insights into the topics that will be covered.

Professor Howe anticipates that students will best be able to help address projects that fall into one of three categories:

1. Prediction. Given a set of data, develop a method to predict how a future data point might behave. Some common examples of these are classification and regression analysis tasks. See more examples.
2. Visualization. What does the data "look like"? Create an interactive web-based visualization designed to illustrate the relationships between a set of variables. E.g., Given sales data, illustrate the most profitable customer segments. See these neat examples.
3. Q&A. What can we find out from the data? Help answer 3-10 difficult, but unambiguous, questions. "How long do visitors typically spend on our website?"; "When we changed advertising campaigns last month, did donations go up or down?"; "What has been the social impact of service A on demographic X?"

Organizations are encouraged to follow these guidelines to ensure their projects are appropriate for the course:

1. Projects should be able to be completed in a short amount of time. The course is only 8-weeks long; it will be important to scope out a small chunk of work to solicit student assistance for.
2. Projects should be well-defined and self-contained. Specialized domain knowledge and the ability to navigate complex legacy software with dependencies are important skills for data scientists to develop. However, given that there is limited time in the course, it is vital to minimize these variables so that organizations can maximize the insights students help provide.
3. The choice of technology to solve the problem should be unconstrained. We won't be able to accommodate specific requests like "We need this to be in Oracle PL/SQL" or "Write a C function to do ...".

For the assignment submission, students will be required to turn in their analysis code, any resulting graphics or visualizations, and a 1-2 page write-up explaining their solution and the insights it helps generate.

When will the analysis be completed by?

All analyses will be completed by the end of the course, which will be towards the middle or end of August.  

How many students will work on my project?

Since this project is part of an optional assignment, it is possible that an organization attempting to recruit students’ help will receive no assistance. Conversely, another organization may receive offers from many interested students. In the case of high demand, it will be up to the organization to manage relationships with students in accordance with their own bandwidth

How can I be sure that students will not share my organization’s data or insights with competitors?

The University of Washington, Coursolve, or any affiliated faculty/staff/researchers cannot be held liable for any data confidentiality breaches provoked by students, and in fact will not ever actually handle your data. All interactions will be between the student and the participating organization. Members of organizations are encouraged to check in with the best practices for nondisclosure and intellectual property agreements laid forth by their entities and act in accordance with these policies.

Will I be able to connect with students after the course is over?

Organizations and students will have the chance to independently determine if they wish to continue collaborating after the end of the course. Course staff or Coursolve will not facilitate interactions with students after the end of the course.  

Who will be taking this course?

The course requires that students have some prior programming experiences, some exposure to databases, and basic understanding of statistics.

Beyond this, there is no way to succinctly characterize the types of students that will enroll. However, data from previous MOOCs suggests that many students will be between the ages of 25-34, have prior work experiences, and be interested in gaining tangible skills to apply in a professional context. Students will be from all over the world, and most will already be credentialed in some sort of undergraduate degree (albeit perhaps not in a field related to data science).

How will students be evaluated?

The optional projects will be evaluated via peer assessment. Students that complete and submit a project for a particular organization will then be tasked with evaluating the submissions of up to five of their peers. These evaluations will be conducted based on a rubric specified by Professor Howe.  More details will be provided later on in the course.