This practicum provides an opportunity for businesses, organizations, instructors, and undergraduate learners to collaborate to identify and translate a real problem into a data analytics problem. Starting with a mission-relevant puzzle or problem from a client, student teams undertake a semester-long project with the client that begins with necessary background domain research, project scope refinement, and a mock bid proposal. Project work typically includes moderate data collection, significant data preparation, descriptive analysis, and statistical modeling/testing - with the potential to include predictive modeling, machine learning implementation, and other more advanced techniques as appropriate to the project and conditional on the prior coursework of students assigned to your project. Practicum project results and recommendations (with other deliverables as agreed to by the client, instructors, and student teams) will be communicated through an executive summary, a formal technical report document, and a live final presentation.

A practicum client will be assigned two student teams of 3-4 undergraduate DA majors in their Junior or Senior year. Each team will come from a different section of the course led by separate instructors. The teams largely work independently in parallel, as a quasi-competition, under the expectation that the client will comparatively identify the strengths and weaknesses of each team's final deliverables and answer some version and/or combination of the following: "Which team would you hire?", "What deliverables hold the most value?", "Which team's suggestions are most like to be implemented?"

Ideal clients are first and foremost interested in partnering with course instructors in the education and preparation of the next generation of well-rounded undergraduate data analytics majors through work on a project of relevance to your organization and/or mission. We like to label an ideal project as "mission curious" rather than "critical" since the course environment is focused on providing an opportunity to undergraduates that are still developing a DA skillset. To ensure course and student learning objectives are achieved, we recommend that the expected database for a project be of sufficient size for some statistical modeling (e.g., at least 5,000+ observations/rows and 20+ variables/columns) in size. Of course, these are just guidelines, and course instructors would be happy to discuss the viability of other data for a successful practicum project. Also, data need not be ‘clean’ in one data file nor entirely collected at the start of the project -- it is advantageous to the students’ learning experience to engage in moderate data collection, preparation, and wrangling before analysis. Ideally, we want to strike a balance so that data prep does not consume student work for the entire project nor present a burden to the client or project viability.

Client projects should keep in mind that most students will have completed course work on introductory data analytics, introductory computer science, applied statistics, and data systems at this stage in their undergraduate DA major. These classes expose students to skills including data cleaning, some SQL, linear and logistic modeling, descriptive statistics, hypothesis testing, reproducibility practices, many data visualization methods. A typical student in the practicum will have experience with Python, R/Rstudio, many typical libraries, GitHub, and Tableau. Some students may have exposure to more diverse software platforms and/or advanced supervised and unsupervised classification and machine learning methods, but this cannot be guaranteed. Thus, successful projects are encouraged, but not limited to, to mesh with these techniques and platforms.

The client-student relationship is at the core of a positive practicum experience. Consequently, clients should be willing and prepared to meet with assigned teams multiple times over the semester, maintain avenues for appropriate communication outside of meetings, offer constructive (and even critical) formative feedback on milestones and performance throughout the semester, and commit to a live (likely virtual) final presentation session at the end of the semester. Some past clients have met weekly with teams, while others relied more heavily upon as-needed electronic communication with 3-5 scheduled meetings spaced across the semester. Instructors generally leave these details up to the client and student teams' needs and comfort.

Potential clients can submit a high-level proposal or problem statement of the project question, scope, and expected deliverables. Information on relevant data sets and their content, units of analysis, and especially availability/readiness is helpful. Any preferences for acceptable tools/formats for the project work ideal deliverables are also valuable at this stage. It is typical for this class to have a project evolve and narrow as clients interact with instructors and especially the students so that the project scope can be organic and multifaceted. We value and recognize the importance of data security and protecting intellectual property. Thus, as needed, datasets could be provided based on a non-disclosure agreement, and/or we can work with you to come up with other secure data sharing solutions. The practicum course instructors will review the documents to confirm the scope and timing of the proposed problem and its alignment with the practicum course requirements.

Some (non-limiting) examples of previous and desirable practicum project topics:

1. Assessment of organization/policy efficacy (e.g., social services, sustainability, food service)
2. Market analysis and viability for new products
3. Consumer trends, personification, and strategy for retail
4. Email campaign strategy
5. Website usage, patterns, efficacy
6. Client/Donor/Consumer engagement and participation
7. Demand for social services (healthcare, emergency services, infrastructure, etc.)
8. Customer acquisition and retention
9. Healthcare and public health models (prior project on medication delivery efficiency w/in a hospital network)
10. Dashboarding and modeling hunger for major regional foodbank
11. Determining media consumption (mass vs. digital)
12. Reduction of client churn (lower abandonment)
13. Cross-sell and upsell opportunities
14. Develop high propensity target markets
15. Customer segmentation (behavioral or transactional)
16. Market Basket Analysis to understand which items are often purchased together