efyang / timelag-graphgen

Generate colored graphs of time-series data

timelag-graphgen

Generate colored graphs/animations of time-series data

Installation

This program uses python 3. To install all of the dependencies, run:

pip install -r requirements.txt --user

Usage

Call python timelag-graphgen/main.py ARGS, where ARGS are

• -i INPUT_FILE (required)
  • The input data file to use
• -f FORMAT (required)
  • The file format that the input data file is in
    • weekss - Single state, weekly (csv format) - don't know if this still works?
      • Start of interval
      • Time
      • Entries_KIN_t0
      • Exits_KIN_t0
      • Entries_FC_t0
      • Exits_FC_t0
      • Entries_CC_t0
      • Exits_CC_t0
    • dayss - Single state, daily - do not use
    • dayssp - Single state, daily, preprocessed - output of preprocess scripts
    • weekssp - Single state, weekly, preprocessed - output of preprocess scripts
    • weekms - Multistate week file (stata format)
      • state
      • county
      • caretype
      • year
      • admitcount
      • exitcount
      • week2
      • week4
      • geo
      • geo1
    • dayms - Multistate day file (csv format, | delimited)
      • date
      • state
      • county
      • caretype
      • admitcount
      • exitcount
      • year
• -s STATE_ID (required)
  • The ID of the state (e.g. CA)
• --coloring COLORING (required)
  • The coloring type to use; can be one of:
    • linear_all_time
    • linear_seasonal
    • discrete_months
    • discrete_months_split_markers
    • discrete_months_polygons
• --limit LIMIT (optional, defaults to autoscale)
  • The maximum limit to use
• --caretype CARETYPE (required)
  • The caretype to view
• --render OUTPUT_FILE (optional, defaults to none)
  • Whether to render to an output file, and if so, where
  • If none, then will show interactive animation
  • If mp4, will render animation to mp4 output file
  • If image (only png tested), will render final step to image file
• --lag LAG_LEN (optional, defaults to 1)
  • The number of lag steps to use, always in the current time unit
  • e.g. for weekly data --lag 1 means a lag of 1 week, for daily data it means lag of 1 day
• --dropyearly (optional, defaults to false)
  • Whether to drop all data more than a year before the current animation time
• --lowess (optional, defaults to false)
  • Whether to use the residuals from lowess smoothing
• --ppf POINTS_PER_FRAME (optional, defaults to false)
  • Points to animate per frame - can vastly help in reducing render times (7 for daily data recommended)

Other useful files

• render_wave_3.sh: script used to render wave 3, might be useful to look over
• preprocess_multi.py and preprocess_mult_daily.py: preprocess multistate files into state-level files
  • Use weekms and dayms files, respectively
  • arg1: input file
  • arg2: output directory