I analyzed the app frame works from the linked article here is the result when the count is larger than one, first column is usage count among popular apps.<p>"cat topmobilesdk.txt | tr "," "\n" | sed -e "s/^ //g" | sort | uniq -c | sort -n"<p><pre><code> 2 Adjust SDK
2 Formatter Kit
2 Google Maps SDK
2 Google+ SDK
2 Hockey SDK
2 Mantle Framework
2 PLCrashReporter
2 Pop Animation Engine
2 PSPDFKit
3 AFNetworking
3 BPXLUUIDHandler
3 FLAnimatedImage
3 Google Conversion Tracking SDK
3 Google Mobile Ads SDK
3 Masonry
3 SocketRocket
3 Tune SDK (Mobile App Tracking)
4 Card.io SDK
4 Cocoa Lumberjack
4 OpenCV
4 TTTAttributedLabel
5 Eigen
6 Bolts Framework
6 Crashlytics
6 Fabric
6 Facebook SDK</code></pre>
In case anyone wants a tabulated view. <a href="https://gist.github.com/mdubbs/41c19c5384ee0386465cff19aa800e8d" rel="nofollow">https://gist.github.com/mdubbs/41c19c5384ee0386465cff19aa800...</a>
Hey there, nice work! I worked on a similar project for Android. Currently, there are about ~3k apps added: <a href="http://www.libtastic.com/" rel="nofollow">http://www.libtastic.com/</a>
We have been doing SDK Recognition for a couple of years here at Apptopia, and it's absolutely not as easy as it's being portrayed here (i.e. run this GIT library). The technology requires constant attention, maintenance and improving otherwise you very quickly find yourself missing analysis for 65-70% of an app's SDKs. In fact, we have recently re-built our SDK Tech completely as we found previous methods (what we call "Binary Analysis") were extremely inaccurate and resulted in over 20% missed SDKs.<p>The comparison of Uber and Lyft is interesting. Our latest recognition for iOS shows Uber with 6 currently installed and 6 others uninstalled since 2013. Lyft made a lot of updates on May 9 and has 8 installed and 12 uninstalled. The only SDK that they are both currently using is SocketRocket.<p>Screenshots from our tool of the specific SDKs recognized:
Uber: <a href="https://www.evernote.com/l/Aro4rGREdBJI_a80JzIRKZ3z2P8BwM3SGrk" rel="nofollow">https://www.evernote.com/l/Aro4rGREdBJI_a80JzIRKZ3z2P8BwM3SG...</a>
Lyft: <a href="https://www.evernote.com/l/ArrWkfAF3OxGWqcsnrCRtgSnRKgSfwH0I0k" rel="nofollow">https://www.evernote.com/l/ArrWkfAF3OxGWqcsnrCRtgSnRKgSfwH0I...</a><p>We offer free trials at www.apptopia.com where you can look at SDK data for over 1 million apps in 196 countries.
This is very similar data to what's available on our site, AppBrain.
If you're interested in more like this, the top ad networks and developer tools are updated daily at <a href="http://www.appbrain.com/stats/libraries" rel="nofollow">http://www.appbrain.com/stats/libraries</a> and each individual app also shows the SDKs it uses at bottom of the app page, eg. Instagram: <a href="http://www.appbrain.com/app/instagram/com.instagram.android" rel="nofollow">http://www.appbrain.com/app/instagram/com.instagram.android</a><p>It seems in the dataset from OP more different SDKs are detected than we do at AppBrain. We invite anyone to report missing SDKs will definitely review the current presented set to add our omissions!
These lists leave out some important frameworks these apps are built on. For example, Pinterest links AsyncDisplayKit. I'm curious how this data is collected. If Keyword matching against known popular public frameworks, it might be worth expanding your queries to include all public CocoaPods.
This is some cool data. Am surprised how often AFNetworking and Crashlytics are used. Unsurprisingly, for advertising, FB and Google Ad SDKs totally dominates. I signed up. When will you be launching the site with full data?
Surprised to see so much use of AFNetworking. Is it really necessary?
Looking at the example code it doesn't seem to add much (these days), with the overhead of managing another sdk.
Is this an indication of anything other than popularity.<p>With programming languages, the general rule seems to be the worst win in the popularity contests (VB, PHP, MySQL come to mind).