A weather app says it’s not going to rain today, so a person leaves their umbrella at home.
20 minutes later, that same person is soaked, standing under an awning, shivering while the rain floods everything in its path. That person is probably annoyed, but it’s not like the rain caused serious problems for them. For someone who owns a delivery company, though, an unreliable forecast can cause actual losses.
That’s why you can’t just use any weather API for your app.
| NOAA’s Global Historical Climatology Network Daily (GHCN-Daily) uses observations from 90,000+ land-based stations worldwide. It draws from approx. 30 different data sources, and quality assurance checks are being used whenever reconstructing/updating the main dataset. – NOAA National Centers for Environmental Information (NCEI) |
You have to know where its data is coming from, how far ahead it forecasts, how often the conditions are updated, and what other details you get with it. Of course, you also have to check the price because, let’s face it, that definitely plays a part.
And when it comes to research/analysis, these same details will matter even more because historical coverage, update frequency, data format, etc., all can directly affect what you do with the dataset.
| Historical weather datasets aren’t necessarily collections of measurements taken at the requested location. The widely used ERA5 reanalysis combines observations from around the globe and numerical weather models via data assimilation in order to reconstruct conditions. – European Centre for Medium-Range Weather Forecasts (ECMWF) |
Read on, and you’ll see a list of the best API providers.
Top Weather APIs
Weather APIs differ in price – of course – but also in data coverage, historical depth, forecast range, available formats, etc.
Let’s go over why these 4 APIs made the list:
Visual Crossing
Let’s say the tool you’re building needs last week’s rainfall, today’s conditions, and next week’s forecast. With some providers, that means you’ll have to work with separate products or endpoints. Visual Crossing has all 3 in its Timeline Weather API, and it covers locations from all around the world with more than 50 years of historical records.
You can download Visual Crossing weather data as either JSON or CSV; you don’t have to start by writing code.
You’ll also see Query Builder on their website. It can help you test locations and date ranges, as well as maps, dashboards, and datasets you can download. All this means you can quickly see what the service returns.
There’s a free plan, of course, and that one includes 1,000 records per day. Paid access starts at $0.0002 (USD) per additional record or $35 per month. Just make sure to watch how the usage is calculated because a 15-day forecast counts as 1 result, but a large historical request can eat up a lot of records.
OpenWeather
This is probably one of the first names you’ll see when you start looking for weather APIs, and that’s because it offers a lot more than your basic forecast.
One Call 3. can show you the next hour minute by minute, the next 48 hours by hour, and 8 days of daily forecasts. It also provides alerts from the government and historical data that goes back more than 47 years. As for the products, they cover air pollution, geocoding, map layers, and bulk historical downloads.
The first 1,000 One Call requests each day are free, and then after that, you’ll pay $0.0015 for each extra call. That sounds simple enough, but keep in mind that the website contains several APIs, subscriptions, and business products. You’ll have to check that you’re actually reading the page for the endpoint you plan to use, so be careful there.
Also, look at its attribution and share-alike rules before you start displaying the data commercially.
Tomorrow.io
Tomorrow.io will tell you it’s going to rain, but it’s more interested in what you plan to do about it.
You can keep an eye on locations, set conditions that are important for whatever you’re doing, and get alerts when those conditions are expected. That’s very useful for someone who’s planning routes for their delivery trucks, running an event outdoors, watching for wind, lightning, and so on. You’ll also find information on air quality, pollen, fires, maritime weather, aviation, and soil moisture, although a lot of specialized material like that is locked unless you have a premium account.
There’s a free plan available, and that gives you 500 requests per day, but keep in mind that they have limits. You get 25 requests per hour and 3 per second. A standard plan covers 120 hours per day and 5 days per month. Premium goes even further than that.
Unfortunately, you won’t know the prices until you ask for them. Prices aren’t public because the cost depends on your usage and datasets.
Weatherbit
If you don’t look closer, you’ll mistake Weatherbit for yet another source for current conditions and forecasts.
If you dig a little deeper, though, you’ll see that there’s also soil moisture, evapotranspiration, lightning, air quality, climate normals, and energy data. Those extras are very valuable, especially for farmers, solar companies, or utility companies. The more information, the better, because the weather is getting more unpredictable, and there are businesses that depend on accurate forecasts.
With a free plan, you’ll get 50 requests per day, current weather, and 7-day forecasts. Keep in mind that you can’t use the API commercially on a free plan! A standard plan costs $45 per month, and you can use it for commercial purposes. It also gives you 16-day daily forecasts, 10 days of hourly forecasts, and minute-level data for the next hour.
If you need historical data, you’ll need to buy one of the higher plans; a Plus plan will provide 5 years of data, while a Business plan gives you 30.
Conclusion
Although there are 5 excellent weather APIs on this list, there may only be 1 or 2 that will work for your particular project.
The point here is to do your research, see what each option offers, and then decide which provider is right for you.
Think about the things that are important to you, like whether your locations are properly covered, whether you get the dates and the forecast range you need, whether you can use the data commercially, etc.
That’ll help you narrow down your options and choose correctly.
