Issues in staging public dashboard

[iantei]

@shankari There are a few issues observed in the staging of public dashboard. Issue 1 and 2 have been fixed, and published as PR. I am not sure how to further debug and understand for Issue 3.

1. [Fixed] Use of expanded_ct_inferred instead of expanded_ct_inferred_u80 to showcase labeled and inferred bar for trips under 80%. Fix the inferred_u80 inferred label bar #159 30635f7

2. [Fixed] Extra information in the plot_title for Replaced modes
   Patch fixes: Eliminate plot_title for Replaced mode stacked bar charts #157 6833854

Extra information in the plot_title for Replaced modes

3. Disparity in the use of imperial vs metric measurement system in plot title and possibly calculation in staging and current main PR. Why is it showing up as "kilometer - kilogram" instead of "miles - lbs" in staging?
   • Update: The issue is observed with both study/program which has use_imperial = "True" like uue which is study-use_imperial, while also observed with ca-ebike and durham which is program-use_imperial
   • There is no issue with study which has use_imperial = False like usaid-laos-ev
   • Pulled the most recent changes from Github em-public-dashboard, and executed with docker-compose.yml configuration as below:

Detailed Diff of docker-compose.yml:

    ashrest2-41625s:em-public-dashboard ashrest2$ git diff docker-compose.yml
diff --git a/docker-compose.yml b/docker-compose.yml
index 38a975a..5c19b05 100644
--- a/docker-compose.yml
+++ b/docker-compose.yml
@@ -22,10 +22,10 @@ services:
     depends_on:
       - db
     environment:
-      - DB_HOST=db
+      - DB_HOST=mongodb://db/openpath_prod_ca_ebike
       - WEB_SERVER_HOST=0.0.0.0
       - CRON_MODE=TRUE
-      - STUDY_CONFIG=stage-program
+      - STUDY_CONFIG=ca-ebike
     ports:
       # ipynb in numbers
       - "47962:8888"

Looks fine in local testing - Main PR.
Note: The issue with Labeled and Labeled and Inferred for ca-ebike is fixed here #155

Tested where?	Charts
In Staging
In Main PR

Could you please help me understand what could possibly be going wrong on [Issue : 3]?

[iantei]

Something interesting:

For usaid-laos-ev, which is study with use_imperial: False

• There is issue in Sensed chart, but not with other charts.

For cortezebikes, which is program with use_imperial: True

• There is no issue in Sensed charts, but with other charts.

Chart Name	Chart
USAID-Laos-ev
Cortezebikes

[iantei]

Testing the above issue in production.

The issue observed above with usaid-laos-ev is also present in the production, while there is no issue with cortezebikes in the production.

Chart Type	Chart
usaid-laos-ev
cortezebikes

[shankari]

Disparity in the use of imperial vs metric measurement system in plot title and possibly calculation in staging and current main PR. Why is it showing up as "kilometer - kilogram" instead of "miles - lbs" in staging?

I am not sure what you mean by this "issue". If it is imperial, we use miles. If it is metric, we use km. That is as expected.

[iantei]

If you look into the chart for ca-ebike:

• For ca-ebike, use_imperial is set to true.
  If you look in the snapshot above for staging in ca-ebike.
  We have the following:
• The text in metric dropdown menu is : "Total trip length (miles) covered by mode"
• The text in the blue section is: "Total trip length (miles) covered by mode All data Default"
• The text in the white section right above the Stacked Bar Chart is: "Kilometers for each mode" [This is not correct]

[shankari]

Ah thank you for the clarification. We discussed this back when you were working on your first PR. If you pass in a GET parameter to the website, it is not and will not be passed to the notebooks. So this is expected behavior.

[iantei]

Yes, I remember that now.

[iantei]

There is an issue #160 (comment) with the current production public dashboard, which I was also able to observe in my local system. The issue is that it uses "Miles" as the hardcoded value instead of choosing between Kilometer/Miles based on the use_imperial.
I have published the PR with the fix - #162