<img src="https://d5nxst8fruw4z.cloudfront.net/atrk.gif?account=J5kSo1IWhd105T" style="display:none" height="1" width="1" alt="">

Nexosis @ Work & Play

Google Trends iPhone Analysis: Know the Story Behind Your Data

November 24, 2017 @ 10:35 AM | Technical, Use Cases


There is a conspiracy theory that when a new iPhone comes out, old models slow down so consumers upgrade their hardware. The theory is farfetched, but slow performance is inevitable for major iOS releases so Jason decided to do an impact analysis using Google Trends data to attempt to uncover real impact.


We have been brainstorming unique use cases for our machine learning API to inspire you to experiment and build cool stuff. Timing was on my side for this latest project. Apple did a major iOS update and almost immediately after a keyboard glitch affected a large number of users that prevented them from typing vowels. I wondered if this was a regular occurrence that correlated with all iOS system releases or new iPhone hardware releases, and if I could pinpoint if iOS releases cause the glitches

After doing some research I found a Harvard study on this exact phenomenon. The study looked at Google Trends data on “iPhone slow” searches and how they correspond to iPhone hardware release dates. Initial analysis found that there is a spike in searches around the release dates, but this just shows that people “feel” that their phones are slower, it doesn’t prove that the phones are actually operating at a slower rate.

Initial analysis found that there is a spike in searches around the release dates, but this just shows that people “feel” that their phones are slower, it doesn’t prove that the phones are actually operating at a slower rate.

The study then compared the “iPhone slow” to “Samsung Galaxy slow” searches over time to determine if this consumer feeling held true for both. The iPhone spikes are on the Apple product release dates, not when Apple announces the new products. This is why it’s crucial you know the story behind your data: It just so happens that every major iPhone hardware release coincides with a major iOS release. If you are a developer you know that this scheduling makes sense and that it is difficult to write perfect software that doesn’t have bugs.

This is why it’s crucial you know the story behind your data: It just so happens that every major iPhone hardware release coincides with a major iOS release. If you are a developer you know that this scheduling makes sense and that it is difficult to write perfect software that doesn’t have bugs.

The study concludes this still just shows correlation and not causation, so I decided to do an impact analysis on the data with our machine learning API to forecast the “iPhone slow” searches had Apple not done major iOS releases. If users with old iPhones waited a few months to before getting the new iOS version, I suspect the spikes in the slow iPhone trend would level out.

In the chart below you can see that without the iOS releases the “iPhone slow” searches would have been smoother overtime, leading me to believe that the the system updates do cause older phones to operate more slowly in various capacities.


iOS Update Trend *The y-axis represents search interest relative to the highest point on the chart for the given region and time. A value of 100 is the peak popularity for the term. A value of 50 means that the term is half as popular.


I did the same for Apple hardware release events.


Apple Hardware Release Trend


Obviously I knew some of the story behind this data prior to this project but similar to what Guy discussed in his "The Joy of Machine Learning" post I learned even more after analyzing it with our API. It shows the value of using fun projects like this to discover more about your data for future use cases. Machine learning allows you to see hypothetical “what if” scenarios to prove or disprove causal impact.


It's your turn - use Google Trends data to experiment with our machine learning API.


Ready to start building machine learning applications?

Get your free API key  Talk to an expert


Jason Montgomery

Jason is the CTO and co-founder of Nexosis. He oversees and participates in engineering, devops, security, data science/machine learning, product, and technical direction. Some claim that he derives his power for good from his mighty beard.