Layout

[fccoelho]

[marfcg]

The report above is nicely accompanied by the interactive map of current and historical situation: ECDC Interactive Atlas Click on "choose data" -> Indicator = influenza "Restrict available data to the following time period:" 2015-W45 to 2016-W12

Here, the levels are: Low: incidence < epidemic threshold Medium: epi. thresh. < incidence < high thresh. High: high thresh. < incidence < very high thresh. Very high: very high thresh. < incidence

The lower right box controls the animation (if checked for time period on "choose data" menu).

The lower left here does not show anything for influenza. It would be nice to show here the incidence curve on top of the channels (weekly expected values and percentiles) as in our report, for the selected State. For the flu, they do not show the incidence for any Country, just the current situation and trend (+ increasing; - decreasing) once you place the cursor above it.

[claudia-codeco]

Great, this interactive atlas is exactly what we need. (Porque estamos falando em inglês? :)

[marfcg]

Porque estamos falando em inglês?

Força do hábito no git e fóruns em geral :P

[fccoelho]

Além disso é legal podermos ter contribuidores de outros países que possam entender nossas mensagens. So let's continue in English! ;-)

On Wed, Aug 31, 2016 at 10:57 AM, Marcelo F C Gomes < notifications@github.com> wrote:

Porque estamos falando em inglês?

Força do hábito no git e fóruns em geral :P

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/FluVigilanciaBR/fludashboard/issues/1#issuecomment-243772393, or mute the thread https://github.com/notifications/unsubscribe-auth/AAIjW1XYC6b6A5jP97T8ka2datEhoSIZks5qlYhPgaJpZM4Jv09H .

Flávio Codeço Coelho

+55(21) 3799-5551 Professor Escola de Matemática Aplicada Fundação Getulio Vargas Praia de Botafogo, 190 sala 312 Rio de Janeiro - RJ 22250-900 Brasil

[xmnlab]

maybe this lib can be used: http://nbviewer.jupyter.org/github/bibmartin/crossfolium/blob/master/examples/MarkerClusterFilter%20example.ipynb

I will take a look :)

[fccoelho]

I dont think we have a need to plot individual flu cases with a marker cluster, @marfcg what do you think?

On Wed, Aug 31, 2016 at 4:26 PM, Ivan notifications@github.com wrote:

maybe this lib can be used: http://nbviewer.jupyter.org/ github/bibmartin/crossfolium/blob/master/examples/ MarkerClusterFilter%20example.ipynb

I will take a look :)

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/FluVigilanciaBR/fludashboard/issues/1#issuecomment-243873330, or mute the thread https://github.com/notifications/unsubscribe-auth/AAIjW02P9xYpcPBPBg7dre22ek_e-6t1ks5qldVlgaJpZM4Jv09H .

Flávio Codeço Coelho

+55(21) 3799-5551 Professor Escola de Matemática Aplicada Fundação Getulio Vargas Praia de Botafogo, 190 sala 312 Rio de Janeiro - RJ 22250-900 Brasil

[marfcg]

I agree with @fccoelho I like ECDC's Atlas layout. Just color scheme for current situation in each region, no marker cluster. Here goes a proposal based on a combination of what ECDC's Atlas show for dengue reported cases, influenza activity and influenza virus detection - sentinel.

• Upper panel remains exactly the same as the flu activity one.
• Bottom panel is enhanced with incidence curve on the left (with background colors based on historical ranges), demographics on the right (age/sex distribution):

[xmnlab]

Great. But would be a problem if we use a centroid marker in each State to receive a click event? I make this question because folium doesn't have this functionality yet (https://github.com/python-visualization/folium/issues/497).

[marfcg]

@xmnlab For that purpose I don't see any problem with using markers. They would need to be small enough to avoid being on top of each other in the northeastern region of Brazil, due to State sizes in that area, though...

[xmnlab]

@marfcg OK!

Another question ... in choropleth map, what is the related value to create the colors range? should be the mean of the chosen SRAG year value for each state?

[marfcg]

@xmnlab For the map, we should present the color based on the activity for the current/chosen week. We compare that value with the thresholds for that state (epidemic, high and very high, specified in files mem-report-2016-[uf,region].csv ). The site should open with the map corresponding to the current situation (i.e., based on last data entry).

When presenting the map for the whole year, we have (at least) two options:

• keep the color corresponding to the incidence peak, or
• use the same procedure adopted by ECDC in their annual report. Which is:
  • Red if the incidence was above the high threshold for at least 5 weeks;
  • Orange if above the high threshold from 1 to 4 weeks;
  • Yellow if crossed the epidemic threshold but not the high one;
  • Green if it did not cross the epidemic threshold.

For the sake of standardization, I'm in favor of adopting the same criteria as of ECDC. We will discuss this matter with the Health Ministry to check what they believe is better for their activities.

[xmnlab]

great! Thank you @marfcg !

[marfcg]

PS: when presenting the map for a specific week, not for the whole year, the color scheme is based on the levels stated on my first comment in this thread:

Here, the levels are: Low: incidence < epidemic threshold Medium: epi. thresh. < incidence < high thresh. High: high thresh. < incidence < very high thresh. Very high: very high thresh. < incidence

So:

• low: green
• medium: yellow
• high: orange
• very high: red

The map should always have a legend stating the corresponding criteria, be it for specific week or a whole year.

[xmnlab]

OK! so, the application should have two time criteria: a whole year or just a specific week.

Is it correct?

[marfcg]

Correct. It'll be an awesome website. Modesty be dammed :D

---

*Sent from mobile device. My apologies for eventual typos. Marcelo Ferreira da Costa Gomes Programa de Computação Científica Fundação Oswaldo Cruz PROCC/FIOCRUZ

On Sep 16, 2016 17:38, "Ivan" notifications@github.com wrote:

OK! so, the application should have two time criteria: a whole year or just a specific week.

Is it correct?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/FluVigilanciaBR/fludashboard/issues/1#issuecomment-247703503, or mute the thread https://github.com/notifications/unsubscribe-auth/AH4S3htUq0MkFsqL3L0MlJ6zVQQbVRmzks5qqv5KgaJpZM4Jv09H .

[xmnlab]

@fccoelho @marfcg

Pelo que entendi o layout do site deveria ser assim (para a visualização por semana):

E para a visualização dos dados do ano inteiro:

Alguma sugestão? Está faltando alguma coisa?

[fccoelho]

Gostei do layout. Me parece completo.

Em 7 de out de 2016 15:32, "Ivan" notifications@github.com escreveu:

@fccoelho https://github.com/fccoelho @marfcg https://github.com/marfcg

Pelo que entendi o layout do site deveria ser assim (para a visualização por semana):

[image: detalhado] https://cloud.githubusercontent.com/assets/5209757/19200956/093b8d08-8ca2-11e6-9d49-d1d5f9e25d36.png

E para a visualização dos dados do ano inteiro:

[image: resumido] https://cloud.githubusercontent.com/assets/5209757/19200962/0f2e4c78-8ca2-11e6-9cc5-40c4d83a1462.png

Alguma sugestão? Está faltando alguma coisa?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/FluVigilanciaBR/fludashboard/issues/1#issuecomment-252327525, or mute the thread https://github.com/notifications/unsubscribe-auth/AAIjW0CUykFhGB8a1HXedMGyw_DLpiP8ks5qxpBCgaJpZM4Jv09H .

[marfcg]

Perfeito!

[xmnlab]

Ok!