Weird result on heatmap

[Beto_Optiart]

I have 3 APs MR33 on a site and we want to analyse people's flow on that área but the result we are having from Location Heatmap is very weird as per attached Picture. Very strange pattern on heatmap

I haven't found any dedicated documentation that should help me trace the cause of the issue. 

Does anyone have an idea?

Tks,

Beto

[PhilipDAth]

What is it doing wrong?

[MerakiDave]

I'm assuming you're clients are more evenly distributed around the full rectangle of that floorplan, and not actually within the triangle as it shows.  A few things could be contributing to that odd result, including the RF environment.  You would also likely get better RF fingerprinting and triangulation results with a 4th AP.  You'd probably also want to double check the basic settings for best & common practices, like eliminating the lowest couple of data rates in each band, use automatic transmit power reduction, etc.  I'd open a case with Meraki Support to take a look at the APs in real time, go over some Dashboard settings and provide some additional insight.  Hope that helps!

[MilesMeraki]

@MerakiDave Has outlined it plane and simple.

1) Ensure the floor plan dimensions are correctly set.

2) Perform an active Wireless site survey to ensure the correct RF channels and powers are being used for optimal RF performance.

3) Try adding a 4th AP placing in the best location identified from the step above.

[MerakiDave]

Thanks @MilesMeraki that's another good point I didn't mention, to make sure your floorplans were properly sized/scaled, while that may not fix the triangular pattern, it's important so the client density heatmap is also properly colored and dots properly distributed.

[The_Dude]

Hey Meraki Dave,

Here is an example of the 'clustering' issue even IF a 4th AP is deployed.

Even if you don't understand Portuguese I think you will get the gist of the issue.

https://community.meraki.com/t5/T%C3%B3picos-em-Portugu%C3%AAs/Ajustes-no-Heatmap/m-p/28706#M94

.

We plan on rolling out hundreds of sites using the raw data from the Meraki scanning API to our proprietary analytics platform.  If we can't get this resolved then we may have to rethink our solution entirely.

Let me know, like, your opinion, man

Cheers

[redsector]

I have got the same issue with more than three or four accesspoints, and with right floorplans sizes.

Its with different accesspoints like MR33, MR42, MR42E, MR52

[The_Dude]

We are also experiencing the same issues on all of our sites.  The issue manifests exactly as depicted with 3 MR33 APs deployed.  Our layouts are scaled and aligned properly and after a site survey it was determined that there was adequate WiFi coverage.  

The issue lies that very few devices are being displayed outside of the 'Bermuda Triangle' formed by the APs.

We have also observed that in our case when any devices are detected outside of the triangle they are usually associated devices.

We are using the scanning API and plotting to our own solution (using the Meraki raw data) using our proprietary heatmap and we are seeing the same issues there as well.  There seems to be an abnormal clustering of detected devices around the APs and along the direct 'line of site' lines between the APs.

The OP is right.  After an exhaustive search of the documentation there doesn't seem to be any explanation or steps to resolve this issue.

Are there any recommended radio settings?  

Do we need to add more APs and if so where?

Here is a document that outlines best practices when using Meraki RTLS that I found AFTER the fact that directly contradicts the marketing collateral that states that NO special physical placement is required and that the existing architecture is sufficient as long as the coverage is adequate.

https://documentation.meraki.com/MR/Monitoring_and_Reporting/Real-Time_Location_Services_(RTLS)

[The_Dude]

Just opened a case with Meraki Support and I even provided the link to this thread for reference.

Here is the response I received:

"I checked internally and verified there is a known issue on our location heatmap page showing incorrect data that is still being worked by our development team. At the moment, we don't have an ETA when will it be fixed but I will let you know if there is an update."

How is this disturbing? 

1) They mentioned 'known' issue and yet after scouring the documentation, community and Reddit forums it was clearly not 'known' but rather highly suspected.

2) They did not make it 'known' in ANY of the documentation.  Instead it seems that this was deliberately hidden or 'veiled' in the RTLS documentation. 

3) This thread was created almost a year ago and the problem still persists.

4) The scanning API is being highly touted by Meraki and there are partners whose solution is predicated on it.  

see: https://apps.meraki.io/details/aislelabs-flow/

5) The suggested resolutions in this thread are smelling to me like 'red herring'.  

I am hoping that by commenting here this raises the priority level of this issue since our proprietary solution is also based on the scanning API.

I also hope that this thread is being monitored by Meraki employees and that this will be properly communicated to the product management team.

Last but not least, I am hoping that the response that I received from the support rep was based on the limited scope of his job function and not the actual situation.

[NolanHerring]

@The_Dude

I recommend your contact your Meraki account rep directly. These forums are not really the right way to escalate an issue like this. I know they are monitored but I honestly do not know to what extent.

[The_Dude]

Hello NolanHerring,

I did contact a rep. regarding this issue and they were a great help in accelerating the process.

You stated:

"These forums are not really the right way to escalate an issue like this. I know they are monitored but I honestly do not know to what extent."

Unfortunately there is a very limited support structure in the country I reside in for Meraki based issues.

I am told that this will change very soon but in the meantime I need to rely on the documentation and community for support. The support responses in the 'Cases' section encourage leveraging the community for support which is exactly what I did.  

Now I don't condone 'Malicious Compliance' in every case but this issue seems to have been around for a while and the suggested resolutions in this thread did not seem to address what could very well be an underlying issue that is affecting multiple users.  The concept of a forum is a many to many approach to discuss issues and ideas that are relevant to the participants.  To what extent the forum is moderated is Meraki's discretion. I chose to post here to elicit responses to build and articulate my case to the support team to have this issue addressed

Someone who is an active and knowledgeable user here with a growing positive reputation posted this excerpt in one of their responses:

" I would reach out to support and see if there are any known issues, the more people that do this, the more pressure they get that its not just a random few people with this issue."

Do you not agree with the above statement?

[NolanHerring]

@The_Dude

I apologize if that came across as crass.

To clarify what I meant was to contact your account rep, not specifically simply 'support'.  I've found that if I reach out to support, what they will do is check to see if there are any other active or known cases with the same or similar symptom for any issue I might be having. If I am lucky, they find something. 

I'm not aware of any method for escalating a ticket with support directly (not like with cisco TAC for example). However, if the issue is critical and a BIG bug, the only way that I know of to 'escalate with Meraki' is via reaching out to my specific account representative.

They usually have the ability/power to get a proper escalation with engineering etc. Which is great for critical issues. Support alone can sometimes drag their feet.

While these forums are an excellent place for assistance from peers, from what I've seen its rare that Meraki support, or a Meraki employee will respond. They do occasionally (like in this thread up above), but these forums are more for support from others like myself that can assist with configuration, or designs (to an extent). You are correct, this is a great place for knowledge transfer. So by all means please keep the discussion going my friend.

I just find that for actual 'firmware' or 'bug' related issues, the forums themselves don't usually ring the gong within Meraki support from what I've seen (could be wrong of course).

What it can do is if you have a problem, and another person and another, then you have a thread where 30 people have the same problem. Then all of those users create support cases, and that hopefully starts a fire within engineering.

I was not aware that you contacted your account rep, so if you did that, then your on the right path 😃

And to your statement: "The concept of a forum is a many to many approach to discuss issues and ideas that are relevant to the participants."

I 100% agree with you and this is indeed the right place to keep the discussions active. Again I simply wanted to make sure that if you reach out to your account rep, you'll usually get more traction that solely relying on the forums. Sorry if that came out wrong.

And I do agree with the last statement you quoted, because I was the one who made it in another thread.   ^_^

[The_Dude]

Thanks Nolanherring! 

Your follow-up response is very helpful and encouraging.

Cheers and Kudos

[wey2go]

I have doubt as to the accuracy of the heatmap. See that the time is in the early hours in the morning (2 am?) and the heatmap is showing that many "Devices" even though not associated so close to the APs. This places is a park and I have placed the APs on the Google Map, so will not have scaling issues (?).

[NolanHerring]

Open spaces like that are much more difficult to do proper location. Requires usually 4 access points or more, and inside a building this means placing them at the corners and ensuring throughout the building they're are enough that the system can run proper algorithms to do a best-guess at device location based on RSSI.

I personally don't really trust/rely on the location heatmap. its kind of like the heat seater in my car in florida. It sometimes gets cold and for those rare moments i might use it. otherwise i forget it even exists 😃

[The_Dude]

Thanks @NolanHerring for the additional info.

I have pretty much hit a brick wall with regards to this issue and even after multiple escalations and support from the account team there doesn't seem to be much hope of a resolution with the current state of the scanning API.

Here's hoping that Cisco has it in their 'DNA' to provide reliable location data for these open 'Spaces' in the next iteration of their API 😉

[whistleblower]

I can´t find the Heat-Map with the Time Lapse anymore? Is this function no longer available in the dashboard?

[PhilipDAth]

It's still there for me under Wireless/Location Heatmap.

[Ameya_Ahir]

Is this behavior fixed with the new location heatmap? We made a few changes on how the heatmap calculates location, so you should have a much accurate view of the heatmap provided you have 3 or more APs for triangulation.