nachtfeuer / covid19 Goto Github PK

• allow using --format multiple times
• creating multiples images on demand

• currently the CSV data is always downloaded and written into the file coiv19.csv
• provide a boolean flag --cache/--no-cache where the default should be --no-cache
• provide --cache-file where the default should be "covid19.csv" being the path and name of file for the downloaded CSV data
• log the options

• Provide standard logging
• Also log Python version and platform
• Also log the data url
• Also log the current options (like width and height)

• Allow --country=all
• Aggegate then cases and deaths per day displaying the data for all countries
• Change titles in graph in "... In All Countries"

• creating an image for each country
• if --format is used as well then create multiple images for each country
• encode the country (lower case) in the filename "covid19-<country>.<format>"

• At the moment the choosen output format is PNG.
• Provide an option with an enum like way offering all output formats that can easily be added (like svg)

Link: https://npgeo-corona-npgeo-de.hub.arcgis.com/datasets/917fc37a709542548cc3be077a786c17_0

It's to verify what data is provided.

Here's one CSV link:
https://opendata.arcgis.com/datasets/917fc37a709542548cc3be077a786c17_0.csv

The Data there is latest only. If you want to have for a range of days you have to use the REST API (not yet fully tested):

https://services7.arcgis.com/mOBPykOjAyBO2ZKk/arcgis/rest/services/RKI_COVID19/FeatureServer/0/query?where=1%3D1&outFields=*&outSR=4326&f=json&dateTime%3ECURRENT_DATE%20-%20INTERVAL%20%2710%27%20DAY

Probably requires investigation in how to use the API:
https://developers.arcgis.com/rest/services-reference/query-feature-service-layer-.htm

• The fitting trendline looks kind of curious because quite a while the starting cases are very low (nothing exponential)
• provide a parameter --initial-cases to allow filtering out those data for visualization
• the shown totals have be the real ones (means: all data count)

• The viewer functionalities are not of interest anymore (zoom, ...)
• Should play nicely with #16
• Each tab should represent one --country=

Depends on implementation (choose what works best):

• either display the image
• or render the graphs

• Currently the window pops up when running the script
• When the purpose is to generate the image only you should be able to run off the window
• Default should be to show the window
• Log the new option

• using one plot method for cases and deaths in a way you could add further plots.
• that plot method plots the concrete data and the trendline.

The current data source has decided not to continue to provide data on daily base.
However there is another data source doing this: https://covid.ourworldindata.org/data/owid-covid-data.csv

The disadvantage that it starts in February 2020 has to be accepted.

• Refactor the code in the main function into a application class
• split section of the main function into readable methods.

Some graphs display a negative y-axis also there cannot be negative values.

• Using the library click to provide command line parameters (width and height).
• Default is 1024x768
• see: https://click.palletsprojects.com/en/7.x/

Since I do not update every day and - secondly - you do not necessarily see the date of last update in the x axis it would be nice to see that information somewhere. Following ideas (try and see what fits best):

• adding "updated: " as subtitle moving current title to left and this subtitle to the right (smaller font and another color)
• adding "updated: " to the legend in brackets
• adding "updated: " to the label of the y-axis

• Show the y ticks on both side of the graphs
• Update the image

• country as filter (default: Germany)
• changing to language "English" for tests to avoid translation files since covid19 data are in English.
• if a language is specified that does not exist exit with error

• show the legend in both graphs at top left.

• The CSV data has four date relvant columns
• The intention is to combine them to one: https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.to_datetime.html

For bigger numbers it would be nice to see the exat value. Following options (try and see what fits best):

• adding subtitle moving current to left and moving new one to the right with smaller font and another color
• labeling last value