Before I started writing this post I thought it would be neat to tally how many weather apps I have tried for iOS, then I realized how many that really was, became sad and gave up.
I have tried a lot of iOS weather apps.
The sad truth is that not a single one is fantastic, some are good, while most border on offensive.
When Dark Sky came up on KickStarter I had an unusual reaction: I was excited. I backed it, and prodded everyone I knew to back it. I wanted the app and I wanted it badly.
When the app came out earlier this week I received a flood of questions about how good it was, but I couldn’t answer because the Puget Sound region of Washington was having an odd bout of very dry weather. The past few days the normal weather has returned and I was able to put Dark Sky through its paces. Here you will find my long winded thoughts on yet another weather app.
The short version: It’s on my home screen and will likely stay there. It’s very good, but not perfect.
A Bit About Needs
My needs for a weather app will vary from your needs. Dark Sky is all about precipitation — not temps or long term forecasts. It is specifically about the type of perception you can expect in the next hour, and when in that hour you will see rain. This is a fantastic tool if you live in the Puget Sound region because 90% of our weather can be split in three categories:
- Overcast (this isn’t “cloudy with blue skies” this is all gray skies all day).
- Cold Rain.
- Not Cold Rain. (I don’t mean warm here.)
The funny thing about it: all three of those weather events visually look the same if you look at the sky. Sure there are ways to tell if it is going to rain, but you can never be sure. Since no self respecting Puget Sounder will be caught with an umbrella, that means we need to know if we need to wear a waterproof jacket or not. Enter Dark Sky.
So for me, Dark Sky is filling a very immediate need.
This is not to say that in other parts of the world Dark Sky won’t be useful — it’s just to say that if you live in Phoenix, AZ you can stop reading.
I’d be lying if I didn’t say that I was disappointed in the icon. For starters it’s blue (luckily there’s a lot of black). It has a border and uses the glossy overlay that Apple puts on by default. I don’t hate the icon, but it’s far from something that I even remotely like.
I can tolerate it, because the rest of the app outweighs its icon.
The heart and soul of this app is the main screen which is the forecast for the next hour. By default Dark Sky pulls up your current location, you can however specify a location if you like. I found that these locations are very sensitive, meaning the data seems to change even if you are only 10-15 miles from the last spot. That alone impressed me.
That’s not all, because Dark Sky does some very important things that are massively helpful:
- Tells you what it thinks the weather is currently doing. This I find is a good baseline for trusting the app. In my testing it is 100% accurate when it thinks the precipitation is light rain or heavier, but 50/50 when it thinks there is only a light sprinkle. I can’t fault that — rain here is very fickle.
- Tells you what to expect in the next hour. Sure lots of apps do this, but here’s how it tells you (examples from what I saw in usage): “Rain in 35 Minutes”; “Possible Sprinkling”; “Light Rain in 3 Min”. I found those little descriptions very helpful — they are the type of forecast that makes the data useable and I love that.
- The last bit on this screen is a little graph. This is the heart of the app — this is the power part of the app. This graph shows precipitation predictions for the next hour, but it does something even more clever: the graph tells you how confident it is in that prediction.
The confidence level in the predictions is the one thing that most apps seem to omit — likely because they pull their weather data from other sources and just display it. Here, Dark Sky is trying to tell you that it sees rain coming, but it’s not sure (or it is really sure, depending on the case). In my testing the app was always confident for the next 30 minutes and usually fairly confident through the hour.
So how does it show the confidence level? Easy. The graphs wiggles and shakes. The more wiggly and shaky the less confident the prediction.
I remember seeing this video about the changes to the graph and I honestly think it was a very clever and intuitive way to design confidence levels into the app.
Overall this is great.
The radar is fantastic looking and it’s actually fast. You can manually move a time slider to animate the radar image and you can zoom in and out on a regional basis.
I don’t use radar much, but I must admit that this radar is so pretty I find myself often just looking at it. (Especially true on the iPad.)
Ok so I haven’t said a negative thing about the app other than the icon — so why isn’t this the perfect app for me? Easy,]: it’s a niche app. It’s great when I want to know about the next hour’s rain — which is 90% of the time — but horrible if I want to long term plan.
I can’t see beyond an hour, I can’t get the temp, wind, or sun forecasts. For everything this app does well, there are loads of things that, by design, the app does not do at all.
Since I was previously using Apple’s Weather app I placed Dark Sky in its spot on the home screen and now just use the widget in notification center for the longer forecasts — this works well for me — but this is a trade-off.
I don’t want all the extra data crammed into Dark Sky, but would it kill them to add at least the rest of the days worth of precipitation information? (e.g. “It gonna rain!”)