Covid incidence map for Germany

[jfexbrayat]

hi all,

I’ve created a notebook to map current covid incidence for Germany, using bokeh and geopandas. The notebook is short and fetches the required data from other sources. More info on my GitHub repository and the notebook can also be run stand-alone on Google Colab.

Feel free to run it yourself, expand and/or adapt to other countries.

Below a screenshot of the final result, showing the window with additional data that pops out when hovering over the map:

[carolyn]

Hi @jfexbrayat ,

Thanks so much for sharing this great work! Your notebook is well organized and commented, and I think it’ll help people working on similar projects.

I’d like to tweet about it on the @bokeh account, if that’s alright with you; do you have a Twitter handle we should tag?

[jfexbrayat]

hey @carolyn

Many thanks for the kind word, and feel free to tweet about it… my handle is @jfexbrayat (like here or on GitHub). That’d be awesome to see other users use this as a skeleton for their own project, or expand this one with additional features.

[odadoun]

@jfexbrayat Good job and thanks for sharing.
In our PyCoa project (www.pycoa.fr) we are trying to parse the maximum national covid db we can (for the moment France, Spain, Portugal, UK, Italy, India, USA, Belgium and recently Germany).
We struggle since few days with the German implementation (we use the Robert Koch Institut’s db) Our problem is from German region coding …

I hope to find my answer in your nb :-). If you have few minutes my issue related to German region is here question concerning RKI DEU db · Issue #110 · coa-project/pycoa · GitHub. By advance thank you very much
Here the actual Pycoa German map

[jfexbrayat]

hi @odadoun

thanks for your interest. I’ve answered on github =)

[odadoun]

thank you so much ! I am progressing

[ghomem]

This is very cool… do you know if the shapes data exists for other countries?

[ghomem]

@jfexbrayat I managed to do something similar based on your idea and code. Thanks

I am just a bit concerned about the lack of performance at each date change. But that is a separate topic.

Online here:

https://coviz.io/#section3

[jfexbrayat]

This is awesome @ghomem!

I guess the ~5s delay when plotting is due to bokeh replotting the map every single time and this may be affected by the server’s specs: is your bokeh server hosted locally or on the cloud? Best would be if bokeh would cache all figures upon starting the server, but I don’t know if it’s possible / how you’d do it.

Anways, pretty cool job!

[ghomem]

Thanks @jfexbrayat !

This runs on a cloud instance with bokeh and nginx. I can always upgrade the instance if nothing better comes up.

I am recalculating the map at each slider change because I don’t have acess to the ColumDataSource that is created inside plot_bokeh. In my non-geo plots that callback of the sliders updates the ColumnDataSource and that makes the plot change, but in this case I don’t have it. I posed the question here:

Performance on Geoplot + how to reach datasource?

Hi, I am using pandas-bokeh to plot geospatial data that has different values every day. The result is here http://localhost:8000/coviz#section3 and works, but it is rather slow and I am not sure how to improve it. One problem might be that plot_bokeh creates a data source internally and we can’t reach it from the oud side. Or can we? The callback for updating the map is creating a new map at each call, which is kind of expensive server side: https://github.com/ghomem/coviz/blob/master/mai…

with a link to the source code.

Comments are welcome.

[ghomem]

@jfexbrayat I solved the problem but it was far from obvious The solution is online now.

Details on the solution are here:

Performance on Geoplot + how to reach datasource?

Just to comment that this problem is solved. For the record: the data source of a plot created with plot_bokeh can be fetched from the plot if we create it with a name https://github.com/ghomem/coviz/blob/master/main.py#L285 it can be kept handy https://github.com/ghomem/coviz/blob/master/main.py#L302 and updated on a callback: https://github.com/ghomem/coviz/blob/master/main.py#L393 Thank you,

[odadoun]

Hi,

sorry I was on vacation I have missed your question …
Yes the shapes exist for other countries but sometimes it is quit difficult to find them.
Please have a look go our geo class :

coa-project/pycoa/blob/main/coa/geo.py

you will find several json description (like FRA,DEU, GBR,IND,USA,MYS, BEL,PRT, ESP, CHL, …)

[odadoun]

This is very very nice !
We plan to do a DashBoard with our soft, like you did ! I think it is better for advertising