The data shown on this site is for weather in the 70122 zip code in Gentilly, New Orleans, LA (near City Park). The data is collected by me, Geoffrey Gauchet, via a custom-built weather station running on a RaspberryPi. Information on how I built it and the customizations I made can be found here.
Here's an explanation of the various datapoints displayed both here and on the @70122weather Twitter account. All temperatures are in Fahrenheit.
Temp ("Ambient Temp" on Twitter)
This is the temperature of the air. This can be affected by sun and other things as well.
This is the temperature of the actual ground. I have a probe shoved into the dirt and this gives the temperature of the earth.
Est. Temp ("Temp" on Twitter)
Because Ambient Temperature can flucuate and be affected by things other than the air temperature, I take the mean of the Ambient and Ground temperatures to derive this temperature that is more realistic. I've calibrated it against known weather readings from DarkSky and the National Weather Service.
For temperatures above or equal to 70°, this represents the Heat Index. For temperatures below 70°, this will represent the Wind Chill. Heat Index take into account the temperature (the Est Temp I derive), humidity, and a variety of constants such as the rate at which sweat evaporates from human skin 6 feet above ground. Here's the equation NWS and I use. Wind Chill is similar, but also factors in wind speed.
This is the relative humdity. This is not Absolute Humidity, which is the ratio of the amount of water vapor to the amount of dry air. The hotter the air, the more moisture it can hold. Relative Humidity is the ratio of current Absolute Humidity to the highest possible absolute humdity for the current temperature. So, 10% Actual Humidity would mean that currently, 10% of the air around you is moisture. However, Relative Humidity of, say, 60%, means that the amount of moisture in the air around is at 60% of whatever the maximum Absolute Humidity for that temperature is. So, let's say the maximum AH for the temperature is 30%. 60% Relative Humidity means that there is 18% moisture in the air (60% of 30 is 18). This is why you can see the RH be 100%, but you are not currently drowning.
As we see above, Relative Humdity tells us how much of the total possible water the air can currently hold is currently water. So 50% humidity means the air is currently holding half of all the water it can hold. Once we hit 100% humidity, that means the air can't hold it any longer and it needs to go somewhere (down to the ground as rain). The temperature when that happens is the Dew Point. A good explanation of why Dew Point matters is here. I calculate the Dew Point using the current temperature and relative humidity, which is a shortcut formula as I don't have expensive equipment that gives me moisture information beyond that.
This is the current Atmospheric Pressure, the force exherted on the surface by the air around it. So, basically how heavy the air is, in a way. Typically measured with a tube filled with mercury, but we're digital here. Measurement would be based on how much the mercury has risen in the tube. Standard pressure at sea level is 1013 millibars (mb). If you follow tropical weather much, you know that lower pressure typically brings wind, clouds, and rain. Higher pressure is usually calm, fair weather.
The compass direction in which the wind is currently blowing. If there is no wind, it'll be represented as "--" instead of a direction such as ENE, S, SW, etc.
Current speed the wind is blowing, in miles per hour. The wind speed is the average sustained winds over a period of time. My weather station samples data from all sensors every 5 seconds for a span of 5 minutes before reporting in the database. For the wind, it takes all of those readings (60 of them) and gives me the average wind speed in those 5 minutes. That value is then reported as the Wind Speed.
This is the maximum, well, gust of wind tha occured in the 5 minute sampling period mentioned above. So, of the 60 readings of wind speed it collects in 5 minutes, the average of all of them is the sustained wind speed, and the highest value in those 5 minutes is reported as a gust.
The rain meter I use has a little bucket that fills up with rain, tips over when it gets full and sends a signal to my weather station and empties the rain and then stands back up empty. This happens when the little bucket has 0.2974mm of rain it. So, I calculate rainfall but counting how many times the bucket tipped and multiply that by 0.2974 (and then convert to inches). I add up every one of those values to get the total rainfall in the 5 minute sampling period. On Twitter, the "Rain Today" value is the total amount of rain that has fallen since 12:00:00am of the current day.
On the Dashboard, the Lightning column will show the number of lightning strikes detected i that 5 minute sampling period, followed by the average distance of those strikes. For example, "3 ~8.28mi" means that in that 5 minute sampling period, 3 lightning strikes were detected and they were an average of 8.28 miles away. The lightning detector is pretty good at filtering out false positives, like manmade signals (my A/C unit kicking on, my neighbor using a drill, etc). It can only detect strikes 40km (about 25mi) away, but environmental factors will sometime sprevent it from detecting things that far away (trees, wind, etc.). It detects electrical pulses in the 500kHz band and only flags those pulses as lightning if the waveform matches that of actual lightning AND after it detects a minimum of 5 flashes within that pulse (When you see lightning, you see it flicker a few times and it's not just one flash). So smaller and far away strikes might not be detected. The distance is also just a radius away from the location of the weather station and does not provide directions. But, using radar images from local weather sources, you can use that to figure out where that lightning strike came from. For instance, if a strike was 6 miles away and the storm is SE of 70122, then you know thats trike was roughly 6 miles SE.
I'm working on using Machine Learning (AI) to forecast the weather using my data. That's a big project, so it'll be a while. Plus, I only have about 2 months of data. I won't be able to get any meaningful forecasts until I have at least a year's worth of data.
The dashboard has some simple graphs and reporting, but I will be adding some more in the future, like being able to see the maximum temperature for a given time period, or average wind speed, biggest wind gust, etc.
If you have any reports or data you'd like to see, let me know! Message me at @70122weather or @animatedGeoff on twitter!