Program Timeline
Table of contents
Project Matching: 08/10/2020 - 09/04/2020
The matching phase takes place between 8/10 and 9/4. From 8/10 to 8/24 we will be collecting student applications and passing them on to Project Partners. On 8/24, the student application will close and Project Partners will have two weeks to wrap up interviews and finalize their student team.
Project Matching Tools
We have created some tools to help students find which project they are best suited for, and to help Project Partners decide which applicants are good fits. Please watch the accompanying video for more information. If project size shows as 0, it means that the partner(s) haven’t specified their expected number of students. If you have any questions or feedback, please reach out to us at ds-discovery@berkeley.edu.
Tool Instruction Student Dashboard Instruction Partner Dashboard Instruction
Workshops: 09/04/2020 - 12/10/2020
Once teams are finalized, students will be asked to complete a series of assignments, attend live working group sessions and dig into asynchronous workshop materials. These resources will culminate in a final presentation to be submitted for the end-of-semester Showcase event. To access live zoom meetings, please visit our bCourses site.
Workshop Description Workgroup Schedule Worksheet Due Personal Goals and Portfolio Management We will work on examining your personal goals and plans and discussing how they can shape the goals of your project and your final deliverable. Many of these projects are very open-ended in their goals, approaches, and deliverables. There are many paths to take in each project engagement. By reflecting on your personal goals, you will be able to approach the project with clearer intentions, clearly communicate with the project partners what you want to accomplish and learn, and choose a project plan that best fits those goals. 09/28， 09/30，10/02 10/08 Defining the project We will discuss how to define the problem, figure out your project goals and milestones, and design a project plan to follow in the coming weeks. Almost all data science projects will change rapidly. Your goals and approaches will change based on what you learn about your data at the exploratory stages, and the results you get at future stages can change your deliverable. Furthermore, every project The planning process is not about creating rigid tracks to follow - it is about planning for contingencies, considering multiple outcomes, and formalizing the assumptions you make about your project and your data. 10/12，10/14，10/16 10/22 Working on a data science team We will discuss ways to split up the data science project workflow among teammates using version control and other project planning tools. We will also discuss communication styles and conflict mediation skills within a team. Data science is different from traditional software engineering projects, where people can often work on features in isolation and there are established tools such as git which make such work easier. With jupyter notebooks (poor git integration), multiple dependencies, and the need to share not only code but also data and assumptions about that data, there are special difficulties associated with working on a data science team. 10/26，10/28，10/30 11/05 Data in context We will discuss issues surrounding data collection/provenance, data generation, and bias. We will integrate discussions of how these issues affect your analysis from an ethical and a results-based standpoint.Unlike many of your classes, real world problems don’t have ready-made, out-of-the-box datasets for your convenience. A majority of data science work involves data cleaning, interrogating assumptions about the data generating process, and ensuring that your data represents what you think it does. Quite frequently, faulty assumptions about your data can tremendously impact your results down the line, as well as the ethical implications of your work. 11/09，11/11，11/13 Final Presentation Communicating results We will share methods of integrating technical and nontechnical explanations to communicate results to diverse audiences. We will discuss how to frame impact and think about results for different types of engagements.Each data science project, along with its impacts, are incredibly unique. Sometimes your results have clear quantitative impacts.Sometimes your explorations have institutional implications. Sometimes your audience might be non-technical. Sometimes you might be aiming to persuade rather than inform. It is important to have a deep box of tools to describe “impact” rather than just a metric in an algorithm. 11/30,12/02,12/04 12/03
Final Showcase: RRR week
In early December, we will be holding a live Zoom event for students to showcase their work to our extended data science community. final submissions will be included on the Division of Computing, Data Science and Society’s website. Submissions due 12/06.