mcyph / global_subnational_covid_data

Normalization and aggregation of subnational covid-19 case statistics for as many countries as possible on admin 1 (ISO 3166-2) and admin 2 levels.

Hi there 👋

🔭 I’m currently working on covid-19 🔢stats collection, 💾archiving and 📈visualization
🌱 I’m currently learning to use pandas, scikit-learn, Keras, ruby on rails, leaflet, android mobile dev with Java, ...
🏛️ Recently I've also been contributing patches to decidim-cfj and other civic-tech projects

My python libraries ⚙️

🐇 speedysvc - fastest, lowest-latency microservices using shared memory
🖌️ cnn_chinese_hw - tolerant Chinese/Japanese handwriting recognition using machine learning
あ char_data - data on Unicode characters/symbols: find out what 🤖, ∈ or Ж are called for example! 🚧
🌏 lang_data - language data: alphabets, translations, number formatting and similar from the Unicode CLDR 🚧
🌍 multi_translit - convert between alphabets: Cyrillic Россия can be converted to Rossiâ, for instance 🚧
🌎 iso_tools - work with language, script and region (ISO 639, 15924 and 3166) codes 🚧
📙 pos_tagger - multi-engine part-of-speech tagging 🚧

My web apps/libraries 🌐

🚦 react_spectrecss_kickstarter - a quick set of UI, chart and map controls for civic tech+hackathons
🐱 langlynx.com - translation software for 100+ languages with dictionaries and encyclopedias
🗣 help-pronounce.com - shows phonetics and pronunciations for words

Other misc stuff ✨

🐍 python_quick_flythru - practical python standard lib docs 🚧

🚧 = in progress/currently being documented

global_subnational_covid_data's Issues

Data from France

Hi,

That's a very nice and ambitious project!

I'm trying to use the data to visualize the evolution of the testing positivity rates in European subnational regions.

I have trouble understanding the data from France in this repo. This file https://github.com/mcyph/global_subnational_covid_data/blob/61aca8779fec173320bef02a2f2016cd3d21761f/casedata/fr/sante_publique/new.csv reports 1357 new cases for region fr-13 and the date 2020-10-03, and this one https://github.com/mcyph/global_subnational_covid_data/blob/61aca8779fec173320bef02a2f2016cd3d21761f/casedata/fr/sante_publique/tests_new.csv reports 14316 new tests for that region/day.

The official data from https://www.data.gouv.fr/fr/datasets/r/406c6a23-e283-4300-9484-54e78c8ae675 has the following lines for region (département) 13 and date 2020-10-03:

13;2020-10-03;5;85;09
13;2020-10-03;22;219;19
13;2020-10-03;42;274;29
13;2020-10-03;23;258;39
13;2020-10-03;22;239;49
13;2020-10-03;30;220;59
13;2020-10-03;20;176;69
13;2020-10-03;9;146;79
13;2020-10-03;8;80;89
13;2020-10-03;7;37;90
13;2020-10-03;188;1734;0

The last row says the number of new cases for 2020-10-03 and region 13 is 188 (number of tests 1734), and the other rows are for particular age ranges.

Does the discrepancy maybe have to do with the handling of the cl_age90 column here? https://github.com/mcyph/covid_19_au_grab/blob/master/overseas/w_europe/fr_data/FRGovData.py

Thank you!