Skip to content

Latest commit

 

History

History
37 lines (36 loc) · 2.36 KB

TODO.md

File metadata and controls

37 lines (36 loc) · 2.36 KB
  1. Brainstorm for aspects of your proposal
  2. Develop metrics that will help inform the domain decisions
  3. Get familiar with the content of potential datasets; understand what might inform your metrics, and what holes there still are
    • Corridor health
    • Buildings
      • Median age of buildings in each corridor
      • Distribution of building age (grouped by 10-year increments)
      • Distribution of building update age (grouped by 1-year increments)
      • Corridor need of renovation: median building/update age -- low (under 5 years), medium (5-15 years), high (over 15 years)
    • Businesses
      • Number of businesses (by category)
      • Age of businesses
      • Share of owners that live within 2 miles (requires a geocoder 🙁)
      • Diversity of businesses: how many business categories have at least (2? 3?) businesses?
    • Engagement
      • Number of visitors, year over year
      • Distance from home block group histogram (for visitors from Philadelphia)
  4. Consider best way to communicate metrics; for example:
    • Should you use time-series graphs? Density/heat-maps? You're certainly not limited in the number of visualizations you can include.
    • Should your report only be at one level of detail, or should you include a break-down by sub-geography (neighborhood, district, etc)?
  5. Write proposal and develop wireframes
    • Include boxes for metrics and sample prose on wireframes
  6. Develop scripts to extract data from sources and load into PostgreSQL and/or BigQuery
  7. Create the structure for your Airflow pipeline and add your extract/load scripts to it
  8. Deploy your pipeline to a cloud server (and document your deployment steps for when -- not if -- you forget them)
  9. Dive deeper into data
    • Experiment and develop queries for metrics, using tools such as PGAdmin, BigQuery, or Jupyter Notebooks
    • Note useful data transformations and queries
  10. Convert explorations into SQL and Python scripts to transform ingested data
  11. Experiment with visualizations of metrics
  12. Create "live mockup(s)" in HTML of dashboard page(s)
  13. Configure a GCS
  14. Convert mockup(s) to template(s)
  15. Create scripts to render template(s) for dashboard page(s)

If there's time:

  1. Add near-by corridors on the corridor detail map