SPC web app feedback / ideas

[dabreegster]

Please list any changes / ideas / feedback. Screenshots, screenshots with scribbled annotations, text description, etc. Preferably chop into individual items.

[HSalat]

To repro:
Checkout the new_schema branch
cargo run --release config/England/rutland.txt
Go to https://alan-turing-institute.github.io/uatk-spc/dev_app/ and feed it data/output/England/2032/rutland.pb
Or you can use a sample https://www.dropbox.com/s/tq6mxtoaiskzips/rutland.pb?dl=0

[griff-rees]

Following @HSalat's instructions to make combined rather than separate issues, some suggestions:

• A button for switching to a logarithmic y-axis for plots (example below)
• A colour blind colour scheme (sadly there is a ticket for addressing this in plotly but I don't think it's resolved, perhaps this is a helpful example where the URL structure includes one of the pallet options in the papers cited (that probably reads quite confusingly but... if you look at the link that should make sense). I guess if the example is enough different colours and you like the style then the URL effectively is enough to copy from?

[HSalat]

C/p from Greg in slack: https://venngage.com/blog/color-blind-friendly-palette/

[ciupava]

from memory of what Hadrien showed yesterday, the side graphs column (bottom left) looks very long, one needs to scroll to see all the graphs - probably more accessible if contained within the page height (not sure how feasible)

[dabreegster]

From live review with @HSalat:

• resizable sidebar
• median age, salaries (keep avg bmi)
• ethnicity, toggle bar chart btwn linear/logarithmic
• SEMI_DETACHED should be -, PRIMARY_SCHOOL primary school
• order bar chart by enum order
• top 20 retail
• when you hide diaries, stuff still showsu p
• flows enabled but choropleth disabled
• enable/disable lockdown data, plumb total_change_per_day
• add new health things
• filter out age > 86
• note about age (removing > 86 bc one category)
• sex assigned at birth, one line text
• ethnicity pie chart, self-reported
• pie chart for nssec8
• pwkstat remove NA and other from chart
• diabetes and bp and cvd, prevalence per age (yaxis percentage of people in that age bucket with cvd)
• include sic, horiz bar chart
• window.alert for json
• pie chart for diaries, breakdown mode choice

[dabreegster]

https://www.ons.gov.uk/census/maps/choropleth/population/household-deprivation/hh-deprivation/household-is-deprived-in-two-dimensions?oa=E00091042 has a nice way of visualizing choropleths -- instead of a big blanket area polygon, just color buildings within those areas.

[HSalat]

Important ++ = Business registry data:

• Total number of jobs available
• Number of jobs by sic as horizontal bar chart
• Histogram of worplace sizes

Then

• Median age & salaries (keep avg for bmi)
• Clean labels + same order as enum
• Filter out age > 86 + note about it (removing > 86 bc one category)
• Note "sex assigned at birth"
• Ethnicity into pie chart + note "self-reported"
• Nssec8 into pie chart
• Pwkstat filter out NA and other from chart
• Diabetes and bp and cvd, prevalence per age (yaxis percentage of people in that age bucket with cvd)
• Transport mode choice into second pie chart
• Ethnicity, toggle bar chart btwn linear/logarithmic

[dabreegster]

Ethnicity, toggle bar chart btwn linear/logarithmic

Ethnicity is now a pie chart, did we want this for something else?

[dabreegster]

The business registry data is not in the SPC output. Are you proposing to add it in a future version, or load the 180MB CSV file in the web app separately?

[HSalat]

For the time being, I would filter the population that has been matched to a workplace and use that data to recover SICs and workplace sizes. I'll try to think of something better in the future e.g. keep basic summary statistics inside about venues in addition to their coordinates.