search

ICPI / DIV

Organizational Material
10 stars 2 forks source link

3rd 90 Dossier Development #133

Closed jb3436 closed 4 years ago

jb3436 commented 5 years ago

Dossier Development

3rd 90 Dossier FY19Q3

Assignees: Anu, Yaa, Mackenzie Provide updates in Analytics Repo All comments and issues for the systems team should be documented in the analytics repo: Single OU pepfar-datim/Analytics#2195 Global pepfar-datim/Analytics#2196

Checklist of Milestones

Requirements :pencil:

Development :chart_with_upwards_trend:

Handover :tada:

Testing (Led by Systems Team) :computer:

Launch :tada:

ICPI Parent Project Management Ticket https://github.com/ICPI/DIV/issues/130

jb3436 commented 5 years ago

Status

Had some issues in Dev - resolved all of those that were discussed. - see GH ticket in analytics repo

Had screenshare session with pano dev team Met with Joseph K for additional tips on how to build in MS On track

jb3436 commented 4 years ago

3rd 90: -next check-in on Tuesday -Anu on TDY this week but will work on outstanding items either during TDY or once she gets back -Yaa to follow-up with polishing items once Anu makes outstanding changes -will accept support from Pooja and Lachandra for non-technical matter

jb3436 commented 4 years ago

Follow Up

  1. P2: Org level should be on top and FY on bottom so that the FYs are next to each other when you have multiple comparators (like you have for p3) a. Adjust formatting for container

  2. P4: a. fine age 2018 should be fine age 2019 (this is on multiple pages) (not dealbreaker—can be done later) i. Discussed w Jacob - agreed to keep 2018 for now. b. graphs are missing secondary % axis. i. Discussed w Jacob - reason is that for stacked bars, it stacks the dots as well - hiding them and using data labels.

  3. Colors incorrect on Selectors and some viz titles (not dealbreaker—can be done later) a. TO Do

  4. Sex colors not according to style guide (not dealbreaker—can be done later) a. TO Do

  5. P5: ref lines needed (not dealbreaker—can be done later) a. TO Do

  6. Notes should be on black, not red (not dealbreaker—can be done later) a. TO Do

  7. P6/7: a. In the style guide we recommend having the D on a bar and the % on a circle to allow the user to distinguish differences in volume (not dealbreaker—can be done later).
    i. TO Do

  8. Please give VLS and VLC separate pages. When I switch to IP the page becomes very crowded. These are good viz and they deserve separate pages so the user can more easily consume. a. Discussed w Jacob - agreed to keep on same page but with a less condensed visual - see below b. We suggest removing the org level selector for this visual, which doesn't add value to the view - it adds clutter

  9. P8: a. ditto of above: please give VLC and VLS own pages so that there is sufficient space when a user selects IM or SNU1. b. Discussed w Jacob - agreed to keep on same page but with a less condensed visual c. Please put Org selector on the upper X axis. Right now I get this: d. To Do

  10. All pages with VLC: Many of these pages have period selectors (and p3 has trends). However, the D for VLC is time referential (TX CURR 2 q ago), which is in reference to the current Q. So if the period selector is set to 18Q4 “TX_CURR 2 Q ago” is still going to = TX_CURR 19Q1. So I would suggest that this just be for the current Q, and not have VLC trends. However, if you worked with DS to create a D metric that is more nuanced (2 q ago with reference to the last q allowed by the filter), that is great and please disregard this comment—I just wanted to confirm. a. Done: Discussed w Jacob - TX metric is correct

  11. P 11: I’m still worried about the axis scale issue here. This issue is compounded by the space that is being taken up by the table—this is a great site-level graph and right now it only has 30% of the real estate on the page (what with text, table, summary graph). I think it would be better to give the site graph more space and set the axis Min to automatic instead of 0.
    a. To Do

  12. P12: a. This should be unsuppressed on the L axis and VLS on the R, please. Those two are related more than unsuppressed and VLC. However, if you wanted to make a site-level graph of VLC gap as well that would also be a good addition. b. Response: displaying unsuppressed and VLS doesn't seem to give us more information, if I know VLS was 90% for a 1000 PVLS, D already know a 100 are unsuppressed, We approached the page as a gap analysis, here are the people not suppressed and the VLC intimates the people proportion of those who got a VL among those eligible. Even when my unsuppressed number is small, if my vlc is low, I know there are a lot more people who may need clinical intervention c. please allow the site-level bar graph to have the full width of the page—there will be a large number of sites (not dealbreaker—can be done later) i. To Do for Fast Follow

  13. P11: You added a quantitative filter on VLC, but users would probably be just as (or more) interested in a quantitative filter on volume or VLS (not dealbreaker—can be done later) a. To Do for Fast Follow

  14. For the site level pages, it would be clearer to drop the Level selector down just above the summary graph so that it is associated with that graph, not the whole page (not dealbreaker—can be done later) To Do for Fast Follow

jb3436 commented 4 years ago

Handed over to systems