#To Do To Do list for everyone

#Group tasks:

• (RC, EB) Formulate questions for Prof. about analytical methods (regression vs. other methods).
• (SG) Work on SQL queries for random selections (contingent on either database or SQL/csv wrapper).
• (All or someone could volunteer to take lead getting started) Start the Presentation outline (https://github.com/stat4701-edav-gps/presentation) - can be edited inline via GitHub website. Edit README.md file.
• (All or someone could volunteer to take lead on scheduling) Figure out which day/time/place for weekend of April 25-26.
• Perhaps we should include some kind of interactive component to the project - discuss at next meeting
• Good idea!

These Other Group tasks will likely need to be broken out in to smaller steps for individual members.

#Individual tasks

##Richie

• Create map that combines distance from roadbed and block height
• Create that map within Crosslet to show the outliers
• Look into some more statistics, especially with k-means on the heights to see if we can improve the regression
• Create the slides

##Emilie

• Write out actual research questions
• Could we use this text from your taxi-analysis repo?
• Put together analysis plan for larger feedback with Richie.
• Draft a list of questions for review with more spatial stats inclined folks.
• Recode and create any new vars
• Complete the basic exploratory work,
• and check assumptions etc. for model fitting.
• Look into big vis package - might be interesting

##Danny

• Create 3D version of distributed building height for census block, like the buffer one but with block.

• Create a percent block covered with building footprint.

• Study sample

• Determine Census Block size range for SQL statement.

• Update that SQL code for BigQuery R code.

• Consider minimum counts of buildings and maybe count of pickups/dropoffs

• New vars:

• Create Census Block flag variable for Park or No Park. didn't do, will be a lot more work. Using Block size and count buildings will help for this.

• Create Census Block flag variable for Buildings or No Buildings.

• Create Census Block Size variable.

• Maybes

  • Maybe also measure of possible maximum distance across block,
  • or measure of block compactness

• Create descriptive roabded stats (pandas) file.

• Merge descriptive describe() stats to Master data file Uploaded to Google BigQuery for R code here.

• New Maps:

• Create Median Distance to Roadbed groupby dataset.

  • map Median Distance

• Database:

• Coordinate with Amazon about possible Free account, setup PostgreSQL (SQLite or equiv. ) database.

• Work with Sankara to get SQL queries to pull data out of database for analysts.

• Alternatively is this doesn’t work we can look into https://code.google.com/p/gcsvsql/

• Posted some BigQuery for R code here

##Sankara