Business analytics · COMM 337 · Python · March 2026

The Google Play market in numbers

A pandas and matplotlib analysis of about 9,700 Google Play apps: which categories are crowded, which ones actually get installed, and whether charging for an app makes sense.

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

Crowded is not the same as popular

Family, Games and Tools have the most apps. But count and reach are different things: the median Game gets a million installs, while the median Medical or Finance app sits at 5,000 to 10,000. Education is small (88 apps) and still has a million-install median and the third-highest average rating.

Median installs by category (log scale). Hover for number of apps and average rating.

Ratings are bunched at the top

The average rating is 4.16 and the median 4.3, with most apps between 4.0 and 4.5. That compression makes rating a weak signal on its own. Its correlation with review count is close to zero (r = 0.07), so a highly rated app is not necessarily a well-known one.

Distribution of app ratings (bins of 0.1).

Free versus paid

Only 7.7% of apps charge. Paid apps rate slightly higher (4.24 against 4.15) but the median paid app has 5,000 installs against 100,000 for free ones. The median paid price is $2.99. Fourteen apps charge more than $300, most of them "I am rich" novelty apps. The business case is on the free side with monetization after install.

Tools

Python: pandas for cleaning (dropped 2,638 rows with missing values), groupby aggregations and correlations; matplotlib for the original charts. Interactive versions here are redrawn from the same dataset.