Google Play Store 2018 Category Analysis
Shiny App | Github
Background
Google Play Store consists of numerous applications within varying categories. The digital distribution service for Android grants users access to millions of apps, in a number of categories, including many family, gaming, and tool, to name a few. There were approximately 2.7 million applications available as of February 2017. The scraped Kaggle data reveals the state of the Google Play Store as of 2018. The dataset highlights 33 categories in the Google Play Store that developers can explore for potential business growth.
Further exploration and analysis can be done on the Shiny app to help organizations make narrow category comparisons and determine which categories contain the largest or smallest number of apps and installations, the characteristics of each category, and more to determine opportunities for app development.
Overview and Ranking
At first glance, we see a left-skewed distribution across all apps in the Google Play Store, or a 4.1 average rating, which indicates a large number of highly-rated apps.
However, since ratings have a low correlation with the number of installs an app receives, less attention should be placed on ratings to determine the number of installs. Instead, the correlation matrix indicates the number of reviews is slightly correlated (at 0.6) with how many installations the app has, and vice versa. This corresponds to the assumption that the number of reviews increases in line with the number of installations.
Top 10 Made
Top 10 Installs
Despite there being more Family apps, as seen in the Top 10 Made bar chart, Top 10 Installs reveal Game apps as the most installed category, followed by Communication and Tools. Therefore, from a business point of view, considering which category is heavily saturated, like Family apps, is something to recognize when thinking about competition and competitors in the market. On the other hand, the bottom 10 apps made can provide an opportunity for growth, with less competition among these categories. In addition, recognizing the categories that are installed the most may provide insight on which categories consumers prefer to download, thus, providing developers or entrepreneurs opportunities to enter this category and compete for installs.
| Category | Number of Apps | Installs |
| FAMILY | 1877 | 6,237,542,505 |
| GAME | 946 | 13,457,924,415 |
| TOOLS | 830 | 8,112,771,915 |
| PRODUCTIVITY | 374 | 5,793,091,369 |
Of the categories that are made and installed the most, Family, Game, Tools, and Productivity fall under both these top 10 lists.
Category Details
Top 3 Category Apps Made: Free or Paid
Taking a look at the top 3 category apps made–Family, Game, and Tools–we see that most of these apps are Free, with a median pay price of $2.99 for paid apps. Together, these 3 categories make up 27.81 billion installations or 36.9% of all installed apps from the Google Play Store.
Top 3 Category Apps Made: Content Rating
Among the top 3 categories of apps, the most common content rating is accessible for Everyone. There is also a notable Teen rating presence between Family and Game apps. In all, having more apps rated as appropriate for everyone can lead to a high number of installations from users.
The two features most commonly found among the top apps made and installed include Free and appropriate for Everyone.
Future Direction
Further expansion of this Shiny app may be done with an up-to-date dataset to determine the characteristics of categories with the most installs. Based on the correlation matrix, high ratings, price, and the number of reviews don't determine the number of Installs. Since most apps are free, installations may be driven by other features of an app not given in the dataset. With an updated dataset, I would like to identify characteristics of categories and apps that lead to an increased number of installs. Categories with the most number of installs not only determine the needs of consumers today but provide an opportunity for developers to enter this competitive market.
The skills I demoed here can be learned through taking Data Science with Machine Learning bootcamp with NYC Data Science Academy.
About Author
