Business analytics · COMM 337 · Python · March 2026

The Google Play market in numbers

An analysis of about 9,700 Google Play apps using pandas and matplotlib. The project looks at which categories are the most crowded, which ones receive the most installs, and whether charging for an app is a viable strategy.

7,021Apps after cleaning
33Categories
92%Free to download
100K vs 5KMedian installs, free vs. paid

Category size and popularity

Family, Games and Tools have the largest number of apps. However, the number of apps in a category does not reflect how often they are installed. The median game has one million installs, while the median medical or finance app has between 5,000 and 10,000. Education is a small category with 88 apps, yet it has a median of one million installs and the third-highest average rating.

Median installs by category on a log scale. Hover for the number of apps and the average rating.

Distribution of ratings

The average rating is 4.16 and the median is 4.3, with most apps rated between 4.0 and 4.5. Because ratings are so concentrated, they are a weak signal on their own. The correlation between rating and number of reviews is close to zero (r = 0.07), which means a highly rated app is not necessarily a widely used one.

Distribution of app ratings in intervals of 0.1.

Free and paid apps

Only 7.7% of apps charge for downloads. Paid apps have slightly higher ratings (4.24 compared with 4.15), but the median paid app has 5,000 installs compared with 100,000 for free apps. The median price of a paid app is $2.99, and 14 apps charge more than $300, most of which are novelty apps. The data suggests that the stronger business model is a free app that earns revenue after installation.

Tools

I used Python, with pandas to clean the data (removing 2,638 rows with missing values), run group-level aggregations and calculate correlations, and matplotlib for the original charts. The interactive charts on this page were redrawn from the same dataset.