Before you spend forever debugging, check if maybe there missing person_address_ids so they aren’t actually 100 each. 
I have no intention of debugging, Kibana is an Angular app and I have no idea. I can create an issue though.
Also Barcelona appears to be in the sea, not very precise
I will try this one too Home · snuids/Elastic-5.0-Country-Map-Visualizer Wiki · GitHub
So @lluismf, I’d like to setup a public demo of the charts you did so far. Is https://github.com/lluismf/openmrs-kibana up to date with the current state of your work?
No it lacks the map, but will update it tonight. If you want more charts, just ask 
1 Like
@janflowers, @pascal, any chance you could suggest more things to report on to keep @lluismf busy and which you would benefit from? 
Bringing this back to your attention I’d like to introduce and welcome @tomhealy, who will be contributing as a PM for this project in his free time. I asked him to try to reach out to implementations to better understand needs. Please stay tuned for any follow-ups from him, thanks!
1 Like
Sorry @raff for the delay - been pulled into a few other things the last few days. I’ll take a look at what our team needs and post some ideas here. Is there a place I can see what @lluismf already has created? I wouldn’t want to suggest things he has already made.
@valvijo Can you take a look here?
@janflowers, I’m working on migrating the work done by Luis to Docker in order to set it up on a demo server. I was hoping to get it done today, but it’s unlikely since I’m only halfway through. I’ll complete it on Monday.
@janflowers and others, the demo is ready at https://goo.gl/6Qa0Dg (I’m sorry it took longer than promised).
A few tips:
- Click the clock in the top-right corner to change the time range you are interested in. It will adjust most of the charts to only include data from the selected period.
- Click on the “Seizure” on the pie chart to drill down the results for all other charts to limit to only that diagnosis.
- Click on any dot on the “Count of patients admitted by month” chart to drill down the dashboard results to that month.
- You can create more charts and metrics from the Visualize tab on the left. They can be then added to the dashboard by clicking the “Add” link in the top menu on the Dashboard screen.
Note the demo is viewable & editable to the whole world so you may see others experimenting with it at the same time.
The link I provided above should however always display the snapshot of the dashboard, which I took.
FYI, I’ve started generating a database with 100,000 patients, which should result in about 4 million observations to test how well Kibana performs with that much data.
1 Like
When I click on Seizure most of the other charts got to 0 or no results. (E.g. Total count of patients becomes 0.)
@darius, strange I thought it used to work, but I think I know how to fix it.
The huge demo db is ready so some time next week I’ll fix that issue when updating the demo to use the huge dataset.
FYI, I’ve already started to load the huge demo db into Kibana. It means Kibana results will be inaccurate until it all completes. I’m measuring the time it will take to load 4 million obs into Kibana.
If you are curious, there’s an instance of refapp at http://devtest04.openmrs.org:8081/openmrs/coreapps/findpatient/findPatient.page?app=coreapps.findPatient, which is running on that huge db. I wonder, if there are any implementations of that size running Reference Application. The patient search, which I linked, is extremely slow. It would be good to prioritize work on performance tests and db tuning of OpenMRS REST and Java APIs.
If MySQL and Elastic live in the same machine should be fast (no network latency).
Yes, 5 seconds to locate “adams” is very slow. Does it use MySQL or Elastic for the search?