Covid analytical dashboard + Bokeh feedback

Hi,

I’ve put together a Bokeh app that provides a detailed visualization for Covid19 data for Portugal in a way that I find be more useful than what is available in the official dashboards. The code is open source and can be reused for similar projects.

It is here:

https://coviz.io/

and here

Major difficulties I had while using Bokeh:

  • understanding how to make tooltips that include the info for all the series and understanding what is the best way to attach them to a specific line
  • dealing with the relationship between a date range slider and the rescaling of the y axis on multiline plots
  • working with dates in general, which was not 100% obvious :slight_smile:

Still, doing visualizations in python is worth any challenge :smiley:

Thanks for Bokeh.

Regards,
Gustavo

3 Likes

Hi @ghomem,

Thanks for sharing, and for the specific feedback! It’s so helpful when our community can suggest areas for improvement.

I get a 500: Internal Server Error when I go to the live site-- is it down?

Hi @carolyn ,

It is up again now at https://coviz.io/.

I have data auto downloading everyday and it sometimes comes with holes, so I will gradually patch the code for better resistance to deviations from the ideal data reporting. This is one such case.

Cheers,
Gustavo

1 Like

Hi @ghomem,

thanks for sharing this very nice dashboard.
I would like to point to you our Python Covid Analysis toolkit (aka pycoa, www.pycoa.fr 100% Open/Free source).
We have added recently Portugal Covid Cases using Direcção Geral de Saúde - Ministério da Saúde Português data.
So far only Covid Cases is available.
You have the possibility to draw map, histo/pie and dateplot chart.
Please do not hesitate to have a look into our code (GitHub - coa-project/pycoa: pycoa Python source code the dev version). There is a demo Nootebook without installation using (the amazing) binder techno
(GitHub - coa-project/coadocker: Dokerfile and click to ‘launch binder’)
Here an example of a map and a histogram.
Hope you will like it.
cheers
Olivier


PS: so far 9 national Covid data are available : PRT, SPA, FRA, ITA, BEL, UK ,DEU, IND, USA (and 2 world DB)

1 Like

Hi @odadoun ,

Thank you for sharing this.

One question: how are you obtaining the Portuguese data?

I am using the repos from the Data Science for Social Good Portugal:

Cheers,
Gustavo

Hi @ghomem,

Thanks for the link, does information on the Portugal Covid variant are present ? Here the csv file I use https://raw.githubusercontent.com/dssg-pt/covid19pt-data/master/data_concelhos_new.csv.
Since this discussion is not related to Bokeh may be we can continue this discussion off line
(using our github account for instance).
Have a good days.
Cheers
Olivier