Is it possible to check Aeris information for a specific PWS?


#1

I have a PWS uploading data directly to WU and a script to move WU data into PWS Weather.
These are the links to my PWS on both systems:

https://www.wunderground.com/personal-weather-station/dashboard?ID=IBUENOSA280
http://www.pwsweather.com/obs/HSMCAB.html

I have a fixed schedule configured in my Rachio device, I have not had rain for the past 3 days but I keep receiving Rain Skip notifications saying that I had 3.81mm of rain in the past 24hs (which is not accurate).

If I go and check on both WU and PWS Weather there’s not a single drop of rain, but it looks like when Aeris is queried it returns with this inaccurate information… and hence my question: Is it possible to check Aeris information for my PWS ?

I appreciate your help in advance.

Diego


#2

@diegoga

I had our support team look into this and the data on PWS weather does match the data we are receiving.

http://www.pwsweather.com/obs/archive/2017/02/HSMCAB.html

Shows precip on Feb 10-14, 18th.

This is the data received from Aeris (if you look for the word precip it matches up days to the site above).

https://jsonblob.com/ead72e20-f7e9-11e6-95c2-310ebe5b9ea3

After seeing this do you still think there is an issue?

:cheers:


#3

I think so. Here’s my rationale.
Aeris shows precip of 3.81mm on Feb 19th and so did a Rain Skip on the 20th (I have a 3mm threshold).

If I look to both my WU or PWS Weather information there was no rain registered on the 19th…

If you look on PWS Weather for the precipitation on the 19th you’ll see there has been none: http://www.pwsweather.com/archivewx.php?id=HSMCAB&d=20170219

Let me know in case I’m reading the BLOB incorrectly.

Tks in advance for your response!
Diego


#4

@diegoga

Something does look odd, we correctly skipped due to an aggregate Feb 18 + Feb 19 which was .44 (.29 + .15) inches of observed precipitation as retrieved from Aeris. So our software is working correctly.

https://jsonblob.com/08eeb691-f89b-11e6-8e23-937e0896a7a7

But you are correct, Feb. 19th on the PWS doesn’t seem to have visible precipitation.

I did notice the retrieval methods were different. EOD24HR versus max24hr

I’ll open a ticket with Aeris and see what they have to say.

:cheers:


#5

Just following up from Aeris.

Hi Franz,

Thanks for reaching out to AerisWeather.

I’ve informed our teams of your findings and they will be investigating today. I will provide their feedback as soon as possible.

Please let me know if you have any additional question in the meantime.


#6

First of all, thank you very much for your help and support on this one.
I wanted to add something I’ve just noticed.

Under device settings I found it had the wrong location. I don’t think that has anything to do with weather evaluation but thought it’s worth to mention. I have the location corrected now.

Regards
Diego


#7

Hey @diegoga-
While this shouldn’t have a huge impact on your day to day weather, it does impact the numbers we would use for our “Seasonal Shift” feature, so it’s probably best you corrected it!
McKynzee :rachio:


#8

@diegoga

Response from Aeris who has amazing customer support.

Franz,

First off, apologies for the delay in response. This was a very unique issue and took our team a bit to find as we wanted to verify if it was a data issue within Aeris, the PWS system or invalid data from the PWS system.

We have found out it was the latter. Reviewing our API ingest logs, it seems the station sent an errant data element from the prior day on the 19th.
Looking here:
http://www.pwsweather.com/archivewx.php?id=HSMCAB&d=20170219&t=day

At 4:15 local time to the station, you will see what looks to be an errant record. Notice that the temperature jumped, as did the humidity, pressure and the errant 24 hour precip.
​​

In reviewing, we found other days where the station sent a single errant record. Such as on Feb 18th at 8:20, 2:20. 2:25:
http://www.pwsweather.com/archivewx.php?id=HSMCAB&d=20170218&t=day

Our team is taking two courses of action:
PWS weather is ignoring the 24 hour precip reading and being errant. Our Data team has already made an adjustment today to the Aeris API so that it can ignore this value as well, marking it with a caution flag. This update will be pushed on Monday or Tuesday of next week.
We are continuing to investigate this station and if we see that it consistently sends errant data, we will add a caution flag to the station itself, which will take it out of the standard search, unless requesting it directly.

I will provide an update next week once we push the update for the summary system and update the previous data. Additionally, I will inform you on the status of the station itself.

:cheers:


#9

@Franz, thank you vey much for your help and support on this one, I really appreciate it.

I agree there are some errant records on PWS Weather (not on Weather Underground though), but I can’t correlate the rain skip with this errant record (or records).

Let’s wait until next week and see the outcome from Aeris actions.

Tks again!

Diego