Mobile Apps: Partial Environmental Analysis

Contents

1. Summary of Results
2. Data Source and Discussion
3. Data Preprocessing
4. Most Common Apps by Genre
5. Most Popular Apps by Genre: App Store
6. Most Popular Apps by Genre: Google Play

<h2>Summary of Results</h2>

From the perspective of gaps in market offerings that a new development company could leverage to launch a successful initial product, we aim in this project to understand trends in popular mobile apps. We will focus on behavior in the Google Play store and Apple iOS App store, and we will limit our attention to apps that are free to download and install. Broadly speaking, we will not account for political, economic, social, or legal factors, hence only performing part of the environmental analysis that such a company should complete.

We will build our work on the assumption that visibility in the respective app stores is essential to success. And, moreover, that visibility is a byproduct of algorithms that favor already popular apps. We will utilize included data on app genre and categorization so that our hypthetical development company may focus on some subset of the app store that is popular with users but also possesses potential for small companies to get their footing. For example, we will find that some categories are entirely dominated by large corporations, against which a small company would never be able to outspend nor outcompute.

Ultimately, we will identify three areas in the Apple store (reference, health and beauty, and catalogs) and five areas in the Google store (health and fitness, house and home, art and design, personalization, and tools) in which a novel idea has the room to rise to a prominent position in the “leaderboards.” The reasoning for including these and exluding other popular categories is included in the course of our discussion.

note: I acknowledge that my hypothetical premise of a new company looking to chase a trend to build an app is almost certainly not a very good recipe for the longevity of the company. If you prefer, please replace this framework for one where an established company is trying to understand the structure of the marketplace and why their existing app is not being recommended often enough.

<h2>Data Source and Discussion</h2>

The data come from two sources: data on the Google Play Store includes 10841 observations and data on the Apple iOS App Store includes 7197 observations.

Each data set possesses its own schema, as can be seen by the given columns:

#setup our standard environment and a dataframe of the dataset
import pandas as pd
import numpy as np

google_raw = pd.read_csv('googleplaystore.csv')
apple_raw = pd.read_csv('AppleStore.csv')

google_raw.columns

Index(['App', 'Category', 'Rating', 'Reviews', 'Size', 'Installs', 'Type',
       'Price', 'Content Rating', 'Genres', 'Last Updated', 'Current Ver',
       'Android Ver'],
      dtype='object')

apple_raw.columns

Index(['id', 'track_name', 'size_bytes', 'currency', 'price',
       'rating_count_tot', 'rating_count_ver', 'user_rating',
       'user_rating_ver', 'ver', 'cont_rating', 'prime_genre',
       'sup_devices.num', 'ipadSc_urls.num', 'lang.num', 'vpp_lic'],
      dtype='object')

As such, our first task is to clean any problem points and designate those columns that will be useful for our present purposes.

<h2>Data Preprocessing</h2>

First, if we inspect the App variable of the Google dataset, we find that there are numerous duplicate apps. More precisely, the number of duplicate apps is:

len(google_raw['App'])-len(google_raw['App'].unique())

1181

If we inspect some of these duplicates:

google_raw[google_raw.duplicated(subset=['App'])]

	App	Category	Rating	Reviews	Size	Installs	Type	Price	Content Rating	Genres	Last Updated	Current Ver	Android Ver
229	Quick PDF Scanner + OCR FREE	BUSINESS	4.2	80805	Varies with device	5,000,000+	Free	0	Everyone	Business	February 26, 2018	Varies with device	4.0.3 and up
236	Box	BUSINESS	4.2	159872	Varies with device	10,000,000+	Free	0	Everyone	Business	July 31, 2018	Varies with device	Varies with device
239	Google My Business	BUSINESS	4.4	70991	Varies with device	5,000,000+	Free	0	Everyone	Business	July 24, 2018	2.19.0.204537701	4.4 and up
256	ZOOM Cloud Meetings	BUSINESS	4.4	31614	37M	10,000,000+	Free	0	Everyone	Business	July 20, 2018	4.1.28165.0716	4.0 and up
261	join.me - Simple Meetings	BUSINESS	4.0	6989	Varies with device	1,000,000+	Free	0	Everyone	Business	July 16, 2018	4.3.0.508	4.4 and up
...	...	...	...	...	...	...	...	...	...	...	...	...	...
10715	FarmersOnly Dating	DATING	3.0	1145	1.4M	100,000+	Free	0	Mature 17+	Dating	February 25, 2016	2.2	4.0 and up
10720	Firefox Focus: The privacy browser	COMMUNICATION	4.4	36981	4.0M	1,000,000+	Free	0	Everyone	Communication	July 6, 2018	5.2	5.0 and up
10730	FP Notebook	MEDICAL	4.5	410	60M	50,000+	Free	0	Everyone	Medical	March 24, 2018	2.1.0.372	4.4 and up
10753	Slickdeals: Coupons & Shopping	SHOPPING	4.5	33599	12M	1,000,000+	Free	0	Everyone	Shopping	July 30, 2018	3.9	4.4 and up
10768	AAFP	MEDICAL	3.8	63	24M	10,000+	Free	0	Everyone	Medical	June 22, 2018	2.3.1	5.0 and up

1181 rows × 13 columns

google_raw[google_raw['App'] == '8 Ball Pool']

	App	Category	Rating	Reviews	Size	Installs	Type	Price	Content Rating	Genres	Last Updated	Current Ver	Android Ver
1675	8 Ball Pool	GAME	4.5	14198297	52M	100,000,000+	Free	0	Everyone	Sports	July 31, 2018	4.0.0	4.0.3 and up
1703	8 Ball Pool	GAME	4.5	14198602	52M	100,000,000+	Free	0	Everyone	Sports	July 31, 2018	4.0.0	4.0.3 and up
1755	8 Ball Pool	GAME	4.5	14200344	52M	100,000,000+	Free	0	Everyone	Sports	July 31, 2018	4.0.0	4.0.3 and up
1844	8 Ball Pool	GAME	4.5	14200550	52M	100,000,000+	Free	0	Everyone	Sports	July 31, 2018	4.0.0	4.0.3 and up
1871	8 Ball Pool	GAME	4.5	14201891	52M	100,000,000+	Free	0	Everyone	Sports	July 31, 2018	4.0.0	4.0.3 and up
1970	8 Ball Pool	GAME	4.5	14201604	52M	100,000,000+	Free	0	Everyone	Sports	July 31, 2018	4.0.0	4.0.3 and up
3953	8 Ball Pool	SPORTS	4.5	14184910	52M	100,000,000+	Free	0	Everyone	Sports	July 31, 2018	4.0.0	4.0.3 and up

google_raw[google_raw['App'] == 'Instagram']

	App	Category	Rating	Reviews	Size	Installs	Type	Price	Content Rating	Genres	Last Updated	Current Ver	Android Ver
2545	Instagram	SOCIAL	4.5	66577313	Varies with device	1,000,000,000+	Free	0	Teen	Social	July 31, 2018	Varies with device	Varies with device
2604	Instagram	SOCIAL	4.5	66577446	Varies with device	1,000,000,000+	Free	0	Teen	Social	July 31, 2018	Varies with device	Varies with device
2611	Instagram	SOCIAL	4.5	66577313	Varies with device	1,000,000,000+	Free	0	Teen	Social	July 31, 2018	Varies with device	Varies with device
3909	Instagram	SOCIAL	4.5	66509917	Varies with device	1,000,000,000+	Free	0	Teen	Social	July 31, 2018	Varies with device	Varies with device

We find that it is predominantly the number of reviews and just simple duplicate entry that accounts for the duplication. As such, we will remove the duplicate observations through the built-in function:

google_raw = google_raw.drop_duplicates(subset=['App'])

#quick sanity check
len(google_raw['App'])-len(google_raw['App'].unique())

0

Next, we will limit our attention to apps in English. We will do so through a function that inspects the frequency of non-ASCII characters and then apply this to our App variable. Note: ultimately, this will remove 30 apps.

def isAscii(string):
    non_ascii = 0
    
    for character in string:
        if ord(character) > 127:
            non_ascii += 1
    
    if non_ascii > 3:
        return False
    else:
        return True

google_raw = google_raw[google_raw['App'].apply(isAscii)]

Our final piece of clean-up for the Google data is to filter out the paid apps, of which there are 753.

google_clean = google_raw[google_raw['Type']== 'Free']

We complete these same steps for the Apple data, without additional comment.

len(apple_raw['track_name'])-len(apple_raw['track_name'].unique())

2

apple_raw = apple_raw.drop_duplicates(subset=['track_name'])

apple_raw = apple_raw[apple_raw['track_name'].apply(isAscii)]

apple_clean = apple_raw[apple_raw['price']== 0]

<h2>Most Common Apps by Genre</h2>

We are interested in identifying the most successful apps on both markets. With our data now in a clean and usable format, we start this analysis by identifying the most common genres for each market. We accomplish this by looking at the frequency occurence of each genre.

#first we can inspect the counts
apple_clean['prime_genre'].value_counts()

Games                1872
Entertainment         254
Photo & Video         160
Education             118
Social Networking     106
Shopping               84
Utilities              81
Sports                 69
Music                  66
Health & Fitness       65
Productivity           56
Lifestyle              51
News                   43
Travel                 40
Finance                36
Weather                28
Food & Drink           26
Reference              18
Business               17
Book                   14
Medical                 6
Navigation              6
Catalogs                4
Name: prime_genre, dtype: int64

#a simple adjustment returns these as percentages of the total
apple_clean['prime_genre'].value_counts(normalize=True)

Games                0.581366
Entertainment        0.078882
Photo & Video        0.049689
Education            0.036646
Social Networking    0.032919
Shopping             0.026087
Utilities            0.025155
Sports               0.021429
Music                0.020497
Health & Fitness     0.020186
Productivity         0.017391
Lifestyle            0.015839
News                 0.013354
Travel               0.012422
Finance              0.011180
Weather              0.008696
Food & Drink         0.008075
Reference            0.005590
Business             0.005280
Book                 0.004348
Medical              0.001863
Navigation           0.001863
Catalogs             0.001242
Name: prime_genre, dtype: float64

Our results indicate that, amongst free English-langauge apps on the App store, games consititute more than half of the total. More generally, apps oriented towards amusement (e.g., games, entertainment, social networking, sports) comprise about 80% of the market, while apps with practical utility (e.g., education, shopping, utilities, productivity, lifestyle) comprise the remaining 20%. However, more work will be needed below as these counts do not indicate the number of users.

For the Google Play store, we find:

google_clean['Category'].value_counts(normalize=True)

FAMILY                 0.184404
GAME                   0.098747
TOOLS                  0.084415
BUSINESS               0.045932
LIFESTYLE              0.039048
PRODUCTIVITY           0.038935
FINANCE                0.037016
MEDICAL                0.035210
SPORTS                 0.033969
PERSONALIZATION        0.033179
COMMUNICATION          0.032389
HEALTH_AND_FITNESS     0.030809
PHOTOGRAPHY            0.029455
NEWS_AND_MAGAZINES     0.027988
SOCIAL                 0.026634
TRAVEL_AND_LOCAL       0.023361
SHOPPING               0.022458
BOOKS_AND_REFERENCE    0.021442
DATING                 0.018621
VIDEO_PLAYERS          0.017831
MAPS_AND_NAVIGATION    0.013994
EDUCATION              0.012865
FOOD_AND_DRINK         0.012414
ENTERTAINMENT          0.011285
LIBRARIES_AND_DEMO     0.009367
AUTO_AND_VEHICLES      0.009254
HOUSE_AND_HOME         0.008351
WEATHER                0.008013
EVENTS                 0.007110
ART_AND_DESIGN         0.006771
PARENTING              0.006546
COMICS                 0.006207
BEAUTY                 0.005981
Name: Category, dtype: float64

indicating a drastically different marketplace. There is a significantly greater emphasis on labelling apps intended for children (i.e., family). Games are significantly less represented. Amusement apps, more generally, occupy either about 22% or 40%, depending on whether we include the family category. And, practical utility apps occupy about 60% of the market.

Overall, because the markets utilize different schemas and the number of apps does not indicate the number of users, we need to extend this analysis.

<h2>Most Popular Apps by Genre: App Store</h2>

One measure of popularity would be the average number of installs for each app genre. Unfortunately, this information is not available in the Apple data, but we can use the number of user ratings as a proxy. Doing so does have a number of hidden variables (e.g.,user enthusiasm, user happiness, user expectations, genre variation), so our results should be taken with some hesitation.

#group by genre, limit results to rating count, generate median
apple_clean.groupby(by="prime_genre")['rating_count_tot'].median().sort_values(ascending=False)

prime_genre
Productivity         8737.5
Navigation           8196.5
Reference            6614.0
Shopping             5936.0
Social Networking    4199.0
Music                3850.0
Health & Fitness     2459.0
Photo & Video        2206.0
Finance              1931.0
Sports               1628.0
Food & Drink         1490.5
Catalogs             1229.0
Entertainment        1197.5
Business             1150.0
Lifestyle            1111.0
Utilities            1110.0
Games                 904.0
Travel                798.5
Education             606.5
Medical               566.5
Book                  421.5
News                  373.0
Weather               289.0
Name: rating_count_tot, dtype: float64

apple_clean.groupby(by="prime_genre")['rating_count_tot'].describe()

	count	mean	std	min	25%	50%	75%	max
prime_genre
Book	14.0	39758.500000	71324.552101	0.0	2.25	421.5	61044.75	252076.0
Business	17.0	7491.117647	10937.996228	11.0	392.00	1150.0	8623.00	38681.0
Catalogs	4.0	4004.000000	6279.618354	213.0	376.50	1229.0	4856.50	13345.0
Education	118.0	7003.983051	21768.784453	0.0	72.25	606.5	3929.25	162701.0
Entertainment	254.0	14029.830709	38978.940379	0.0	190.50	1197.5	7534.00	308844.0
Finance	36.0	31467.944444	59291.260242	0.0	72.00	1931.0	26411.75	233270.0
Food & Drink	26.0	33333.923077	79117.344084	1.0	60.25	1490.5	9406.25	303856.0
Games	1872.0	22812.924679	94903.404955	0.0	127.75	904.0	7424.50	2130805.0
Health & Fitness	65.0	23298.015385	79689.893021	0.0	115.00	2459.0	7754.00	507706.0
Lifestyle	51.0	16485.764706	53313.761068	0.0	261.00	1111.0	6171.00	342969.0
Medical	6.0	612.000000	664.210509	0.0	7.25	566.5	1174.50	1341.0
Music	66.0	57326.530303	182381.869800	0.0	569.25	3850.0	23445.25	1126879.0
Navigation	6.0	86090.333333	140543.142497	5.0	1035.75	8196.5	119386.00	345046.0
News	43.0	21248.023256	59704.048306	0.0	113.50	373.0	9933.50	354058.0
Photo & Video	160.0	28441.543750	174228.470086	0.0	249.75	2206.0	11960.50	2161558.0
Productivity	56.0	21028.410714	35409.954611	0.0	647.75	8737.5	22278.25	161065.0
Reference	18.0	74942.111111	232099.622506	0.0	729.00	6614.0	18210.50	985920.0
Shopping	84.0	26919.690476	60002.319755	0.0	582.25	5936.0	19590.25	417779.0
Social Networking	106.0	71548.349057	310475.511080	0.0	434.75	4199.0	28539.75	2974676.0
Sports	69.0	23008.898551	50891.091415	0.0	131.00	1628.0	16979.00	290996.0
Travel	40.0	28243.800000	80065.894662	0.0	185.25	798.5	13284.00	446185.0
Utilities	81.0	18684.456790	58816.596859	0.0	83.00	1110.0	8873.00	479440.0
Weather	28.0	52279.892857	107252.562636	0.0	24.75	289.0	46570.00	495626.0

A box and whisker plot is a little easier to see the above (because of the difference in scale I won’t plot all of them)

apple_clean[apple_clean['prime_genre']== 'Catalogs'].boxplot(grid=False,column='rating_count_tot')

<AxesSubplot:>

But, if we focus on one of the genres with a large standard deviation, it becomes rather ineffective.

apple_clean[apple_clean['prime_genre']== 'Music'].boxplot(grid=False,column='rating_count_tot')

<AxesSubplot:>

This is because there are substantial outliers in 6 genres (Music, Navigation, Photo & Video, Reference, Social Networking, and Weather). It would be helpful to identify which apps are the outliers for these genres.

#filter to chosen genre, limit results to two columns, sort the values, and limit results to the top five
apple_clean[apple_clean['prime_genre'] == 'Music'][['track_name','rating_count_tot']].sort_values(by=['rating_count_tot'], ascending=False).head()

	track_name	rating_count_tot
4	Pandora - Music & Radio	1126879
8	Spotify Music	878563
35	Shazam - Discover music, artists, videos & lyrics	402925
60	iHeartRadio – Free Music & Radio Stations	293228
151	SoundCloud - Music & Audio	135744

apple_clean[apple_clean['prime_genre'] == 'Navigation'][['track_name','rating_count_tot']].sort_values(by=['rating_count_tot'], ascending=False).head()

	track_name	rating_count_tot
49	Waze - GPS Navigation, Maps & Real-time Traffic	345046
130	Google Maps - Navigation & Transit	154911
881	Geocaching®	12811
1633	CoPilot GPS – Car Navigation & Offline Maps	3582
3987	ImmobilienScout24: Real Estate Search in Germany	187

apple_clean[apple_clean['prime_genre'] == 'Photo & Video'][['track_name','rating_count_tot']].sort_values(by=['rating_count_tot'], ascending=False).head()

	track_name	rating_count_tot
1	Instagram	2161558
54	Snapchat	323905
65	YouTube - Watch Videos, Music, and Live Streams	278166
166	Pic Collage - Picture Editor & Photo Collage M...	123433
167	Funimate video editor: add cool effects to videos	123268

apple_clean[apple_clean['prime_genre'] == 'Reference'][['track_name','rating_count_tot']].sort_values(by=['rating_count_tot'], ascending=False).head()

	track_name	rating_count_tot
6	Bible	985920
90	Dictionary.com Dictionary & Thesaurus	200047
335	Dictionary.com Dictionary & Thesaurus for iPad	54175
551	Google Translate	26786
715	Muslim Pro: Ramadan 2017 Prayer Times, Azan, Q...	18418

apple_clean[apple_clean['prime_genre'] == 'Social Networking'][['track_name','rating_count_tot']].sort_values(by=['rating_count_tot'], ascending=False).head()

	track_name	rating_count_tot
0	Facebook	2974676
5	Pinterest	1061624
43	Skype for iPhone	373519
48	Messenger	351466
51	Tumblr	334293

apple_clean[apple_clean['prime_genre'] == 'Weather'][['track_name','rating_count_tot']].sort_values(by=['rating_count_tot'], ascending=False).head()

	track_name	rating_count_tot
22	The Weather Channel: Forecast, Radar & Alerts	495626
89	The Weather Channel App for iPad – best local ...	208648
95	WeatherBug - Local Weather, Radar, Maps, Alerts	188583
133	MyRadar NOAA Weather Radar Forecast	150158
138	AccuWeather - Weather for Life	144214

We can infer from these results that new applications trying to enter the marketplace via one of these genres will have a difficult time pulling attention away from the juggernauts, with the possible exception of reference, which we discuss below. The genres besides these six may provide a more reasonable path to notoriety via recommentation algorithms that leverage “most popular” type metrics, and we inspect them below.

If we inspect the reference genre:

apple_clean[apple_clean['prime_genre'] == 'Reference'][['track_name','rating_count_tot']].sort_values(by=['rating_count_tot'], ascending=False).head(15)

	track_name	rating_count_tot
6	Bible	985920
90	Dictionary.com Dictionary & Thesaurus	200047
335	Dictionary.com Dictionary & Thesaurus for iPad	54175
551	Google Translate	26786
715	Muslim Pro: Ramadan 2017 Prayer Times, Azan, Q...	18418
738	New Furniture Mods - Pocket Wiki & Game Tools ...	17588
757	Merriam-Webster Dictionary	16849
913	Night Sky	12122
1106	City Maps for Minecraft PE - The Best Maps for...	8535
1451	LUCKY BLOCK MOD ™ for Minecraft PC Edition - T...	4693
2280	GUNS MODS for Minecraft PC Edition - Mods Tools	1497
2766	Guides for Pokémon GO - Pokemon GO News and Ch...	826
2844	WWDC	762
2895	Horror Maps for Minecraft PE - Download The Sc...	718
5721	VPN Express	14

We see that, while this genre does contain outliers, they are unique in the service they provide and do not serve the broader needs of the category. For the most part, once one moves beyond the top five apps, the remaining applications focus on reference features for popular media. This creates some potential for a new app to exist in this space without worrying about pulling users away from the christian bible or dictionaries. Secondly, based on the title offerings, there seems to be heavy correlation here with the popularity of the family genre, giving some indication of what topics have the potential to be successful within this genre. Finally, it is reasonable to assume that such apps allow for more on-screen time and there are reasonable methods that can be employed to increase user engagement (e.g., tip of the day, rare tidbits, top 10 secrets). All of the above in conjunction with our initial results in this section establishing the popularity of reference may result in an advantage for new developers to enter this space. Contrast this to, say, weather, which is popular but is also dominated by outliers, has very little latitude for a unique value proposition, and lends itself to quick checks.

Next, we see that productivity is rather evenly distributed and completely dominated by large tech companies, making it a very poor choice for a small developement company.

apple_clean[apple_clean['prime_genre'] == 'Productivity'][['track_name','rating_count_tot']].sort_values(by=['rating_count_tot'], ascending=False).head(15)

	track_name	rating_count_tot
123	Evernote - stay organized	161065
150	Gmail - email by Google: secure, fast & organized	135962
168	iTranslate - Language Translator & Dictionary	123215
186	Yahoo Mail - Keeps You Organized!	113709
291	Google Docs	64259
306	Google Drive - free online storage	59255
352	Dropbox	49578
361	Microsoft Word	47999
412	Microsoft OneNote	39638
475	Microsoft Outlook - email and calendar	32807
481	Hotspot Shield Free VPN Proxy & Wi-Fi Privacy	32499
514	Documents 6 - File manager, PDF reader and bro...	29110
585	Google Sheets	24602
590	Microsoft Excel	24430
644	Inbox by Gmail	21561

In shopping and food & drink, we observe that popular apps are predominantly tethered to an external store, making them also poor choices for our developers.

apple_clean[apple_clean['prime_genre'] == 'Shopping'][['track_name','rating_count_tot']].sort_values(by=['rating_count_tot'], ascending=False).head(15)

	track_name	rating_count_tot
30	Groupon - Deals, Coupons & Discount Shopping App	417779
71	eBay: Best App to Buy, Sell, Save! Online Shop...	262241
142	Wish - Shopping Made Fun	141960
157	shopkick - Shopping Rewards & Discounts	130823
162	Amazon App: shop, scan, compare, and read reviews	126312
198	Target	108131
212	Zappos: shop shoes & clothes, fast free shipping	103655
235	Walgreens – Pharmacy, Photo, Coupons and Shopping	88885
251	Best Buy	80424
274	Walmart: Free 2-Day Shipping,* Easy Store Shop...	70286
314	OfferUp - Buy. Sell. Simple.	57348
326	Apple Store	55171
329	Shop Savvy Barcode Scanner - Price Compare & D...	54630
382	Ibotta: Cash Back App, Grocery Coupons & Shopping	44313
422	letgo: Buy & Sell Second Hand Stuff	38424

Sports is also, as is to be expected, dominated by the large brands. It is very unlikely that an individual developer would be able to keep pace with the data collection and analysis that these organizations are able to perform, making this a poor choice.

apple_clean[apple_clean['prime_genre'] == 'Sports'][['track_name','rating_count_tot']].sort_values(by=['rating_count_tot'], ascending=False).head(15)

	track_name	rating_count_tot
62	ESPN: Get scores, news, alerts & watch live sp...	290996
94	Yahoo Fantasy Sports	190670
125	WatchESPN	159735
135	The Masters Tournament	148160
147	Yahoo Sports - Teams, Scores, News & Highlights	137951
289	ESPN Fantasy Football Baseball Basketball Hockey	64925
305	CBS Sports App - Sports Scores, News, Stats, W...	59639
313	FOX Sports Mobile	57500
334	2016 U.S. Open Golf Championship	54192
368	NBC Sports	47172
386	NBA	43682
411	ESPN Tournament Challenge	39642
428	2016 US Open Tennis Championships	37522
547	NFL	27317
637	MLB.com At Bat	21830

Health & Fitness does provide reasonable opportunity. It is fairly evenly distributed at the top and provides a wide array of functionality to the user. There is potential for a unique idea to establish itself within such a context.

apple_clean[apple_clean['prime_genre'] == 'Health & Fitness'][['track_name','rating_count_tot']].sort_values(by=['rating_count_tot'], ascending=False).head(15)

	track_name	rating_count_tot
20	Calorie Counter & Diet Tracker by MyFitnessPal	507706
42	Lose It! – Weight Loss Program and Calorie Cou...	373835
149	Weight Watchers	136833
209	Sleep Cycle alarm clock	104539
231	Fitbit	90496
338	Period Tracker Lite	53620
463	Nike+ Training Club - Workouts & Fitness Plans	33969
545	Plant Nanny - Water Reminder with Cute Plants	27421
758	Sworkit - Custom Workouts for Exercise & Fitness	16819
860	Clue Period Tracker: Period & Ovulation Tracker	13436
880	Headspace	12819
929	Fooducate - Lose Weight, Eat Healthy,Get Motiv...	11875
1000	Runtastic Running, Jogging and Walking Tracker	10298
1059	WebMD for iPad	9142
1089	8fit - Workouts, meal plans and personal trainer	8730

Finally, catalogs might provide some opportunity, under the right circumstances. Because there are so few apps, despite it being a less popular genre, there is some room for a free app to be noteworthy. However, such apps seem to require either substantial technical implemenation or a large offering of options, both of which will substantially increase the development time.

apple_clean[apple_clean['prime_genre'] == 'Catalogs'][['track_name','rating_count_tot']].sort_values(by=['rating_count_tot'], ascending=False).head(15)

	track_name	rating_count_tot
862	CPlus for Craigslist app - mobile classifieds	13345
2028	DRAGONS MODS FREE for Minecraft PC Game Edition	2027
3276	Face Swap and Copy Free – Switch & Fusion Face...	431
3891	Ringtone Remixes - Marimba Remix Ringtones	213

Ultimately, we have identified three areas (Reference, Health & Beauty, and Catalogs) where new developers may aim to gain visibility within the marketplace.

Closing Caveat: there is, of course, the idea that any developer is able to establish themselves with a succificiently unique value proposition, and the above analysis does not prohibit such a possibility. We are predominantly concerned with how they can establish visibility within each genre.

<h2>Most Popular Apps by Genre: Google Play</h2>

The Google data contain information on the number of installs (which seems like a reasonable metric on popularity, but could be improved if we knew the rate of duplicate installation by users), but we can see that it doesn’t provide fine detail:

google_clean['Installs'].unique()

array(['10,000+', '500,000+', '5,000,000+', '50,000,000+', '100,000+',
       '50,000+', '1,000,000+', '10,000,000+', '5,000+', '100,000,000+',
       '1,000,000,000+', '1,000+', '500,000,000+', '500+', '100+', '50+',
       '10+', '1+', '5+', '0+'], dtype=object)

Because we aren’t able to establish ordinality there will be room for improvement with our results, but these results will allow us to identify any trends within the most popular apps as a single category. For such purposes, we will identify each category with its minimum (e.g., associate [10k,50k) to 10k) and convert to a numeric data type to allow for computations.

google_clean.iloc[:,5] = google_clean.iloc[:,5].apply(lambda x: x.replace(',', '').replace('+', ''))

google_clean.iloc[:,5] = pd.to_numeric(google_clean.iloc[:,5], downcast="float")

Now we can continue as we did with the Apple data.

google_clean.groupby(by="Category")['Installs'].median().sort_values(ascending=False)

Category
ENTERTAINMENT          3000000.0
WEATHER                1000000.0
VIDEO_PLAYERS          1000000.0
EDUCATION              1000000.0
SHOPPING               1000000.0
GAME                   1000000.0
PHOTOGRAPHY            1000000.0
COMMUNICATION           500000.0
FOOD_AND_DRINK          500000.0
HEALTH_AND_FITNESS      500000.0
HOUSE_AND_HOME          500000.0
ART_AND_DESIGN          100000.0
TRAVEL_AND_LOCAL        100000.0
MAPS_AND_NAVIGATION     100000.0
AUTO_AND_VEHICLES       100000.0
PARENTING               100000.0
PERSONALIZATION         100000.0
PRODUCTIVITY            100000.0
SOCIAL                  100000.0
FAMILY                  100000.0
COMICS                  100000.0
SPORTS                  100000.0
TOOLS                   100000.0
BEAUTY                   50000.0
BOOKS_AND_REFERENCE      50000.0
NEWS_AND_MAGAZINES       50000.0
DATING                   10000.0
FINANCE                  10000.0
LIBRARIES_AND_DEMO       10000.0
LIFESTYLE                10000.0
MEDICAL                   1000.0
BUSINESS                  1000.0
EVENTS                    1000.0
Name: Installs, dtype: float32

Based on some of the reasoning we did with Apple data, there are a few categories we can quickly exclude: Weather (as discussed above), video players (technical competition, as discussed above), education (domain expertise), shopping and food/drink (brick-and-mortar requirement, as discussed above), photography (large tech domination, domain expertise), sports (as discussed above).

While Entertainment is very popular, it is dominated by very large streaming services (which would be far too difficult for a small company to try and tackle) and apps that are oriented towards children. It is theoretically possible for a small company to get a foothold in the latter consideration, but it would require substantial investments in art, design, and domain expertise in children’s entertainment. As such, this would only be a fruitful path for our hypothetical company if they wanted to focus on media for children and build out the relevant departments.

google_clean[google_clean['Category'] == 'ENTERTAINMENT'][['App','Installs']].sort_values(by=['Installs'], ascending=False).head(15)

	App	Installs
865	Google Play Games	1.000000e+09
855	Netflix	1.000000e+08
888	IMDb Movies & TV	1.000000e+08
893	Talking Ben the Dog	1.000000e+08
874	Talking Angela	1.000000e+08
866	Hotstar	1.000000e+08
889	Twitch: Livestream Multiplayer Games & Esports	5.000000e+07
879	Talking Ginger 2	5.000000e+07
892	PlayStation App	5.000000e+07
886	Amazon Prime Video	5.000000e+07
859	YouTube Kids	5.000000e+07
953	HBO GO: Stream with TV Package	1.000000e+07
955	PlayKids - Educational cartoons and games for ...	1.000000e+07
1002	SketchBook - draw and paint	1.000000e+07
1000	Imgur: Find funny GIFs, memes & watch viral vi...	1.000000e+07

We can see that the game category has a rather parallel difficulty to the children’s entertainment apps. Unless our hypothetical development company seeks to work in exclusivly game development and populate functional art and design departments, opportunities are limited here.

google_clean[google_clean['Category'] == 'GAME'][['App','Installs']].sort_values(by=['Installs'], ascending=False).head(15)

	App	Installs
1654	Subway Surfers	1.000000e+09
1655	Candy Crush Saga	5.000000e+08
1722	My Talking Tom	5.000000e+08
1661	Temple Run 2	5.000000e+08
1662	Pou	5.000000e+08
1758	Hungry Shark Evolution	1.000000e+08
5950	Banana Kong	1.000000e+08
6554	Skater Boy	1.000000e+08
9166	Modern Combat 5: eSports FPS	1.000000e+08
1781	Trivia Crack	1.000000e+08
1773	Extreme Car Driving Simulator	1.000000e+08
1764	Pokémon GO	1.000000e+08
1763	Piano Tiles 2™	1.000000e+08
1653	ROBLOX	1.000000e+08
1743	Hill Climb Racing 2	1.000000e+08

Health and Fitness, in contrast to the Apple data, is rather homogeneous, consisting of various iterations of tracking applications. This may indicate a gap that could be filled, or it could point to a substantial difference in how the two markets operate. We will mark this as a potential category, but we will need to do future work that inspects if the difference in categorization accounts for the difference in app variety between the two data sets.

google_clean[google_clean['Category'] == 'HEALTH_AND_FITNESS'][['App','Installs']].sort_values(by=['Installs'], ascending=False).head(15)

	App	Installs
5596	Samsung Health	500000000.0
1360	Period Tracker - Period Calendar Ovulation Tra...	100000000.0
1286	Calorie Counter - MyFitnessPal	50000000.0
1256	Home Workout - No Equipment	10000000.0
1344	Headspace: Meditation & Mindfulness	10000000.0
1283	Garmin Connect™	10000000.0
1361	Period Tracker Clue: Period and Ovulation Tracker	10000000.0
1357	Period Tracker	10000000.0
1289	Endomondo - Running & Walking	10000000.0
1292	Runkeeper - GPS Track Run Walk	10000000.0
1277	Runtastic Running App & Mile Tracker	10000000.0
1296	8fit Workouts & Meal Planner	10000000.0
1317	Google Fit - Fitness Tracking	10000000.0
1316	Daily Workouts - Exercise Fitness Routine Trainer	10000000.0
1312	Nike Training Club - Workouts & Fitness Plans	10000000.0

House and Home contains some potential in the design aspects (I assume rental/sales aggregators would be difficult to gain traction against). If we combine this with our observations from the reference category in the apple data, there may be room here to focus on or partner with the correct brand to create a focused app with a pre-existing user base.

google_clean[google_clean['Category'] == 'HOUSE_AND_HOME'][['App','Installs']].sort_values(by=['Installs'], ascending=False).head(15)

	App	Installs
1446	Zillow: Find Houses for Sale & Apartments for ...	10000000.0
1454	Trulia Real Estate & Rentals	10000000.0
1449	Realtor.com Real Estate: Homes for Sale and Rent	10000000.0
4205	tinyCam Monitor FREE	10000000.0
1456	Houzz Interior Design Ideas	10000000.0
1470	DaBang - Rental Homes in Korea	5000000.0
1465	Trulia Rent Apartments & Homes	5000000.0
1482	Alfred Home Security Camera	5000000.0
1450	Real Estate sale & rent Trovit	5000000.0
1453	CYANOGEN. Rent, buy an apartment, a room, a co...	1000000.0
9984	Room Creator Interior Design	1000000.0
1479	Apartments.com Rental Search	1000000.0
1491	ColorSnap® Visualizer	1000000.0
1476	Rent.com Apartment Homes	1000000.0
1494	Room Painting Ideas	1000000.0

Art and Design certainly is dominated by technical implementations that would be hard for a small company to compete with. But, like House and Home above, apps like Tattoo Name and Garden Coloring indicate that niche subjects may be able to get a foothold in this category. Of course, how to identify the correct niche so that the dice roll is guaranteed to be a success is beyond my expertise.

google_clean[google_clean['Category'] == 'ART_AND_DESIGN'][['App','Installs']].sort_values(by=['Installs'], ascending=False).head(15)

	App	Installs
3	Sketch - Draw & Paint	50000000.0
42	Textgram - write on photos	10000000.0
45	Canva: Poster, banner, card maker & graphic de...	10000000.0
19	ibis Paint X	10000000.0
12	Tattoo Name On My Photo Editor	10000000.0
2	U Launcher Lite – FREE Live Cool Themes, Hide ...	5000000.0
18	FlipaClip - Cartoon animation	5000000.0
37	Floor Plan Creator	5000000.0
7	Infinite Painter	1000000.0
8	Garden Coloring Book	1000000.0
10	Text on Photo - Fonteee	1000000.0
11	Name Art Photo Editor - Focus n Filters	1000000.0
22	Superheroes Wallpapers | 4K Backgrounds	500000.0
16	Photo Designer - Write your name with shapes	500000.0
1	Coloring book moana	500000.0

Travel and local would be very difficult for a small company to break into. Certainly, there may be local apps that could gain traction in a focused locality. But, by definition, these could never be large, break-out applications.

google_clean[google_clean['Category'] == 'TRAVEL_AND_LOCAL'][['App','Installs']].sort_values(by=['Installs'], ascending=False).head(15)

	App	Installs
3127	Google Street View	1.000000e+09
3117	Maps - Navigate & Explore	1.000000e+09
3121	Google Earth	1.000000e+08
3112	Booking.com Travel Deals	1.000000e+08
3115	TripAdvisor Hotels Flights Restaurants Attract...	1.000000e+08
3103	trivago: Hotels & Travel	5.000000e+07
3125	VZ Navigator	5.000000e+07
9833	MAPS.ME – Offline Map and Travel Navigation	5.000000e+07
3151	2GIS: directory & navigator	5.000000e+07
3130	Goibibo - Flight Hotel Bus Car IRCTC Booking App	1.000000e+07
3142	Foursquare Swarm: Check In	1.000000e+07
3143	PagesJaunes - local search	1.000000e+07
3144	Flightradar24 Flight Tracker	1.000000e+07
3145	Yatra - Flights, Hotels, Bus, Trains & Cabs	1.000000e+07
3149	Despegar.com Hotels and Flights	1.000000e+07

The peronalization and tools categories appear to be rather haphazard grab-bags of add-on features. Some of these would require large catalogs to compete against (e.g., Backgrounds HD). The antivirus applications would provide the same technical domination that we have seen numerous times above. But, given the correct novel add-on feature, there does seem to be room here to gain some traction. Hence, we will mark them as possibilities.

google_clean[google_clean['Category'] == 'PERSONALIZATION'][['App','Installs']].sort_values(by=['Installs'], ascending=False).head(15)

	App	Installs
3446	GO Keyboard - Emoticon keyboard, Free Theme, GIF	100000000.0
4474	Parallel Space - Multiple accounts & Two face	100000000.0
3354	ZEDGE™ Ringtones & Wallpapers	100000000.0
3425	Backgrounds HD (Wallpapers)	100000000.0
3385	Hola Launcher- Theme,Wallpaper	100000000.0
3360	CM Launcher 3D - Theme, Wallpapers, Efficient	100000000.0
3374	APUS Launcher - Theme, Wallpaper, Hide Apps	100000000.0
4812	GO Launcher - 3D parallax Themes & HD Wallpapers	100000000.0
3400	Ringtone Maker	50000000.0
3428	Koi Free Live Wallpaper	50000000.0
3382	Yandex Browser with Protect	50000000.0
3365	ZenUI Launcher	50000000.0
3352	Nova Launcher	50000000.0
3443	iKeyboard - emoji, emoticons	10000000.0
3444	Simple Neon Blue Future Tech Keyboard Theme	10000000.0

google_clean[google_clean['Category'] == 'TOOLS'][['App','Installs']].sort_values(by=['Installs'], ascending=False).head(15)

	App	Installs
3234	Google	1.000000e+09
3265	Gboard - the Google Keyboard	5.000000e+08
7536	Security Master - Antivirus, VPN, AppLock, Boo...	5.000000e+08
3235	Google Translate	5.000000e+08
4005	Clean Master- Space Cleaner & Antivirus	5.000000e+08
3255	SHAREit - Transfer & Share	5.000000e+08
4808	Avast Mobile Security 2018 - Antivirus & App Lock	1.000000e+08
3266	Google Korean Input	1.000000e+08
3333	Speedtest by Ookla	1.000000e+08
4151	Google Now Launcher	1.000000e+08
3272	Share Music & Transfer Files - Xender	1.000000e+08
5077	AppLock	1.000000e+08
8896	DU Battery Saver - Battery Charger & Battery Life	1.000000e+08
4080	Lookout Security & Antivirus	1.000000e+08
5695	AVG AntiVirus 2018 for Android Security	1.000000e+08

Thus, we have identified five categories in which a novel idea may be able to gain visibility: health and fitness, house and home, art and design, personlization, and tools