#goals THE STATISTICIAN AS PROJECT MANAGER: STRATEGIES FOR SUCCESS & SERENITY Jennifer Thompson, MPH Vanderbilt University Medical Center Department of Biostatistics Critical Illness, Brain Dysfunction & Survivorship Center bit.ly/jlt-wsds2018 · @jent103
LIFE OF AN ACADEMIC BIOSTATISTICIAN Study Data Develop Interpret, Descriptives, design, collection, research inference, disseminate analysis cleaning, prediction question results strategy management Statistician Statistician Statistician Statistician Statistician What does “missing” mean? What happened when…? Clinical support Principal PIs PIs(?) team investigators Infrastructure team bit.ly/jlt-wsds2018 · @jent103
LIFE OF A CORPORATE DATA SCIENTIST Data Interpret, Develop Develop Descriptives, disseminate acquisition, business analysis inference, results; cleaning, deploy, monitor prediction question strategy management models Data scientist Data scientist Data scientist Data scientist Data scientist What does “missing” mean? What happened when…? Domain Business Stakeholders Stakeholders teammates (?) stakeholders Infrastructure team bit.ly/jlt-wsds2018 · @jent103
when there is no project manager… We become a project manager bit.ly/jlt-wsds2018 · @jent103
…AND THIS HAS DRAWBACKS ADVANTAGES! • Project management is a lot Deeper project/domain • knowledge of work… • …which is not typically Stronger relationships with • considered for career teammates advancement Better understanding of raw data • • Doing this well may actually These lead to stronger analysis, • hold us back, if it takes time interpretation, communication from “promote-able” tasks & skills development Many of us enjoy it! • Link: Tanya Reilly's excellent talk on "glue work” bit.ly/jlt-wsds2018 · @jent103
strategies for success that have worked in my context 爛 YMMV : take what works, tweak what doesn’t bit.ly/jlt-wsds2018 · @jent103
COMMUNICATION Regular team meetings Public list of priorities • Group discussion of goals, • Trello board, Slack channel… projects, timelines means • Gives stakeholders a clear everyone understands picture of when their projects rationale will be addressed • Facilitates updates on • Facilitates conversation about roadblocks, successes shifting priorities as needed • Allows all team members to learn from each other bit.ly/jlt-wsds2018 · @jent103
FOCUS 🔏 bit.ly/jlt-wsds2018 · @jent103
KEEP THE MISSION IN MIND What are your team’s goals? 🎰 What projects/analyses will move you toward them? Limit your to-do list to those projects! bit.ly/jlt-wsds2018 · @jent103
DEALING WITH (SCOPE) CREEPS Problem : Scope creep (n, my version) : Project gets bigger and bigger without explicit intent 😉 One solution : Use your project/statistical analysis plan as a tool 🛡 • Record original intent V1.1 V2 V1 • Document & version Added plan for Outcomes totally Original plan handling missing changed every change data • Additions/changes mean updating the version; this might be okay, but process forces you to think bit.ly/jlt-wsds2018 · @jent103
FOCUS VS AD HOC PROJECTS MONDAY 臘 FRIDAY “I'm going to work on this big Spent my week on 18 clinical trial! It's going to be revisions, bug fixes, “can you awesome!” just tell me how many…” Results : 1. Group is only inching toward goals, rather than making major progress 2. I am slowly being driven mad bit.ly/jlt-wsds2018 · @jent103
FOCUS VS AD HOC PROJECTS Our specifics: Goal : • Two weeks working on major project MAJOR (eg, primary outcomes of clinical trial) • One week on smaller needs Make progress (reviewer responses, bug fixes…) on major work • Repeat! 🔅 that moves us MAJOR toward primary Results: mission, while • Major projects done more efficiently remaining • Smaller projects done in reasonable responsive to time frame, communicated well how many… • Everyone’s needs handled also-important can you tell me… reviews are back… • I feel more satisfied with tasks & output ad hoc requests can you change… bit.ly/jlt-wsds2018 · @jent103
PROCESSES ✅ bit.ly/jlt-wsds2018 · @jent103
WHO MAKES DECISIONS? Data scientist Stakeholders Data engineers Additional stakeholders C-suite Consulting data scientist bit.ly/jlt-wsds2018 · @jent103
WHO IS… Task: Developing an SAP R esponsible Biostatistician Research fellow A ccountable Biostatistician more: bit.ly/dpm-raci C onsulted Faculty mentor I nformed Additional faculty, research coordinator, DB team bit.ly/jlt-wsds2018 · @jent103
DOCUMENTATION Project request form for gathering initial info In progress: Onboarding • Goals for analysis docs for new team members • A ccountable & c onsulted team members Deadlines ⏱ What’s already available? 📞 • Logistics Funding source 💹 Target product 🎰 Data source(s) 💼 IRB permissions 🔒 Potential roadblocks discussed before data work gets underway; clearer vision of decision-makers, end product, & value added bit.ly/jlt-wsds2018 · @jent103
SYSTEMIC CONSIDERATIONS bit.ly/jlt-wsds2018 · @jent103
• Can we do a better job at recognizing these professional skills? Relevant: Manisha Desai’s WSDS2017 talk on recognizing team scientists in academia Varied career ladders - industry currently does • this better than academia IMO Update typical promotion criteria • If this work is valuable, provide training/support • • Offset time spent on PM tasks by asking for protected time to either perform those tasks, or to focus on promote-able tasks professional development, developing group infrastructure… • Are you a manager/supervisor ? Do you see this happening? Step in! • Know what is valued by your organization, your career path - and yourself bit.ly/jlt-wsds2018 · @jent103
ACKNOWLEDGEMENTS & RESOURCES • VUMC Strategy & Innovation Office for project management training • Jesse Mostipak & Sharla Gelfand for nonprofit & industry perspectives 👰 Related reading! Trey Causey on the data scientist as product manager (product != project, but many similar concepts) Tanya Reilly on “glue work” (h/t to Sharla!) Roger Peng on balancing resources, analysts managing the flow of information bit.ly/jlt-wsds2018 · @jent103 · jenthompson.me
Recommend
More recommend
