This course will become read-only in the near future. Tell us at community.p2pu.org if that is a problem.

Data Evaluation [Nov. 11, 2012, 2:52 p.m.]



  • 2 Types of Data Needs:
  • One Time Data Need: This is something you can do at the Data Dive. There is no need to worry about reproducibility or the organization's access to tools or know-how.
  • Recurring Data Need: A recurring data need would involve data that is continually updated or requires new analysis depending on changing conditions (this could be daily, monthly or even yearly). When making recommendations about recurring data needs the most important thing is to make sure any process you suggest is realistic to implement by the organization

 

  • 2 Data Sources
  • Internally Produced: Data produced as a byproduct of normal operations (e.g. Intake surveys, server logs, financial information) or data produced as part of regular reporting (e.g. Annual Survey). This may be a place that you can recommend structures and organization schemes that may be helpful (we will talk more about that in the second Data Jam)
  • Externally Produced: Data produced by other organizations that is often publicly available. This would include anything from state or local information to Census Data and beyond.
  Internally Produced Data Externally Produced Data
One Time Data Need
  • Use whatever tools you like
  • Still document your methodology
  • Consider that it may become recurring
  • Think about how their collection impacts your analysis (random samples)
  • Research!
  • Use whatever tools you like
  • Document sources, tools and methodology
Recurring Data Need
  • Think about the organization and collection of data. Does it lend itself to analysis?
  • What are the costs and benefits of gathering more data?
  • Focus on creating a reproducible process rather than on actual data
  • Be aware of what resources are available to the organization
  • Think about how this process can be updated