weather station data collection

the goal is long term data collection and web posting

to play, I have a DS18B20 read every minute
averaged the readings for the hour
takes the highest reading in the hour
takes the lowest reading for the hour

posts to the database each hour. currently to google sheets

so, that is 3 readings, posted, high low and average.

has anyone run a weather station for some time and found something needed or wanted ?

part of this is also for degree days, both heating and cooling
and to trend for spring planting for crops, mid season and end of season.

Is there wifi or lan at the site? How are you planning to post this to the website?

The data I'm reading is measured inside the house: temperature and humidity. I pull the same things externally from my local airport and Darksky. I serve up graphs of it, so I can see the high & low without a need to calculate or store it.

I recently started storing barometric pressure, but haven't done anything with it.

Rainfall measurement would be nice. Maybe air quality too.

full WiFi is used to send data once per hour to Google Sheets for graphing and storage.

since I want to set this up for long term, years of data, I don't want to save every minute for fear of too much data.

Once per hour seems like an acceptable amount of posted data.
the high/low/average seems like it will help offer more details and keep the count acceptable.
degree day is done with a reading a minute so the 60 minute average is posted as well.

seems like there are a lot of 'how to' instructions and not so much of 'what I learned over the years...' or 'what mistakes I made..'

Weather data is great but doesn’t really make a lot of sense until you gather it from other locations (microclimate).
Build a weather station and have others in your neighborhood use it too. Then, you’ll start to see very interesting data. I know that’s not really answering your question, but even something like temperature in a 1 mile radius is very interesting.

czu001:
Weather data is great but doesn’t really make a lot of sense until you gather it from other locations (microclimate).
Build a weather station and have others in your neighborhood use it too. Then, you’ll start to see very interesting data. I know that’s not really answering your question, but even something like temperature in a 1 mile radius is very interesting.

the area will be rather large
I have a friend at work who wants one for his yard and others have expressed their interest.
I would also like to make sure that after a few years, the data has some historic value as well.
how much to record and how often is something one would not really appreciate the value of until some time, and then is it too late to get more.
also, there was a guy on here last year who wanted to use the impending weather to ramp up the temperature of his floor. since floor heating has such a huge thermal time off-set, one could get a cold snap, start heating, then about the time the floor recovered, the sun came out. so using something as open as GoogleSheets is attractive as well.

Hourly data is probably enough for history and trends. I have noticed however that it makes my 24 hour charts look strange or fake. I capture data every ten minutes and since it then goes to a MySQL database, the data volume hasn’t been a problem so far.

If you are graphing the data, it might be worth sampling with greater granularity even if you don’t keep it all long term.

There does not seem much point in collecting long-term weather data unless it is collected in accordance with the standards used by the official meteorological agencies.

Local weather data (such as the fact that there was frost on the grass outside my door at 7am a few mornings ago) is just a local curiosity unless it can be integrated with the international view of the earth's climate.

If your data is collected according to the standards then your local meteorological agency would probably be happy to have the data to supplement their own recording stations.

I'm in the UK and the daily forecast by the UK Met Office seems to me to capture the local conditions sufficiently accurately for my needs. I am mostly surprised by how accurate it is. I can plan my outdoor activities to avoid rain (when it is not continuous) and I think the forecast would also be sufficient for forward planning for a slow-reacting heating system.

...R

Robin2:
There does not seem much point in collecting long-term weather data unless it is collected in accordance with the standards used by the official meteorological agencies.

Local weather data (such as the fact that there was frost on the grass outside my door at 7am a few mornings ago) is just a local curiosity unless it can be integrated with the international view of the earth's climate.

If your data is collected according to the standards then your local meteorological agency would probably be happy to have the data to supplement their own recording stations.

I'm in the UK and the daily forecast by the UK Met Office seems to me to capture the local conditions sufficiently accurately for my needs. I am mostly surprised by how accurate it is. I can plan my outdoor activities to avoid rain (when it is not continuous) and I think the forecast would also be sufficient for forward planning for a slow-reacting heating system.

...R

I cannot find were to get historical data.
forecasts can be found, but for plants the year to year cycles change over time and cycle over periods.
what was the results of the last season or the year before
if it was dreadfully hot, what was the duration and how did the plants fare ?
I am more worried about trends.
which is why I am looking at hourly to start.
I will have a look to see standards from our meteorological agency(s)

dave-in-nj:
I cannot find were to get historical data.

AFAIK all the meteorological agencies have historical data for the past 100 years or more. Of course they may charge for access to it.

...R

I would suggest using something like AWS to host your data. It's extremely cheap (free) if you're under their usage limits, and it sounds like you will be well under it. I'm not sure how comfortable you are with programming backend, but you could use API Gateway to host your service, RDS to store your data (it's a big step up from Google Sheets), and Lambda to process the API request and stuff it into RDS.

Then, you could use a Lambda function to run every 4-6 hours to average your data into a separate table. Once your data is in the database, you could export it out into your Google Sheets for graphing.

dave-in-nj:
I cannot find were to get historical data.

For the US historical station data is available here: Local Climatological Data (LCD) | Data Tools | Climate Data Online (CDO) | National Climatic Data Center (NCDC)

cactus.io

in particular: