How to Use Python to Analyze the Fitness Tracker Market: A Step-by-Step EDA Guide
The global market for fitness trackers and smartwatches has exploded in recent years, as more and more health-conscious consumers embrace these wearable devices to monitor their activity levels, sleep, heart rate, and other biometrics. According to a 2023 report from Grand View Research, the global fitness tracker market size was valued at $38.2 billion in 2022 and is expected to expand at a compound annual growth rate (CAGR) of 15.2% from 2023 to 2030.
In India specifically, the fitness tracker market has seen a major boom, with shipments of smartwatches growing 364.1% YoY to 6.3 million units in the first quarter of 2022 alone, per Counterpoint Research. Indian consumers are increasingly opting for budget smartwatches over traditional wristwatches and fitness bands.
With such tremendous growth and competition in the space, it‘s more important than ever for fitness tracker brands, retailers, and enthusiasts to leverage data to their advantage. Python is an ideal programming language for this purpose, thanks to its powerful libraries for web scraping, data manipulation, and visualization.
In this step-by-step guide, we‘ll walk through how to use Python to scrape fitness tracker product data from an e-commerce website, preprocess that data, and perform exploratory data analysis (EDA) to derive valuable insights into the market landscape, consumer preferences, and competitive trends. Whether you‘re a data scientist, business analyst, or hobbyist, these techniques will help you unlock the power of data to better understand the booming fitness tracker industry.
Step 1: Web Scraping Fitness Tracker Data
The first step in any data science project is gathering the data. In this case, since fitness trackers are predominantly sold through e-commerce channels, we can use web scraping to extract product information from a site like Amazon, Flipkart, or a brand‘s own online store.
Python offers several libraries to make web scraping easier, including BeautifulSoup, Scrapy, Selenium, and Requests-HTML. For this example, we‘ll use Scrapy to scrape fitness tracker listings from Amazon.in.
First, install Scrapy using pip:
pip install scrapy
Next, create a new Scrapy project and Spider:
scrapy startproject amazon_trackers cd amazon_trackers scrapy genspider trackers amazon.in
Open the newly created trackers.py file in the spiders directory and edit the parse() method to find and extract the desired data from the HTML using CSS selectors:
def parse(self, response):
for tracker in response.css(‘div.s-result-item‘):
yield {
‘brand‘: tracker.css(‘h5.s-line-clamp-1::text‘).get(),
‘model‘: tracker.css(‘h2.a-size-mini::text‘).get(),
‘price‘: tracker.css(‘span.a-price-whole::text‘).get(),
‘rating‘: tracker.css(‘span.a-icon-alt::text‘).get(),
‘reviews‘: tracker.css(‘span.a-size-base::text‘).get(),
}
next_page = response.css(‘a.s-pagination-item::attr(href)‘).get()
if next_page is not None:
yield response.follow(next_page, callback=self.parse)
This code snippet will scrape the brand, model name, price, average rating, and number of reviews for each fitness tracker listing, and follow pagination links to crawl subsequent pages.
Finally, run the spider to start scraping:
scrapy crawl trackers -O trackers_dataset.csv
This will save the scraped data to a CSV file named trackers_dataset.csv. Repeat this process for other e-commerce sites or marketplaces to gather a more comprehensive dataset.
Step 2: Data Cleaning and Preprocessing
With our raw fitness tracker data in hand, the next step is to clean and preprocess it to ensure data quality and consistency. This typically involves handling missing values, removing duplicates, converting data types, and normalizing or standardizing numerical features.
Load the scraped data into a Pandas DataFrame:
import pandas as pd
df = pd.read_csv(‘trackers_dataset.csv‘)
print(df.head())
print(df.info())
Check for missing data:
print(df.isnull().sum())
Assuming there aren‘t too many missing values, we can simply drop those rows:
df = df.dropna()
Remove any duplicate listings:
df = df.drop_duplicates(subset=[‘brand‘, ‘model‘])
Convert prices from strings to floats:
df[‘price‘] = df[‘price‘].str.replace(‘,‘, ‘‘).astype(float)
Extract the numeric ratings from the text strings:
df[‘rating‘] = df[‘rating‘].str.split().str[0].astype(float)
Remove non-numeric characters and convert the number of reviews to integers:
df[‘reviews‘] = df[‘reviews‘].str.replace(‘,‘, ‘‘).str.extract(‘(\d+)‘).astype(int)
With these basic data cleaning steps complete, our fitness tracker dataset is ready for exploratory analysis. Of course, depending on the specific data you‘re working with, additional preprocessing may be necessary, such as encoding categorical variables, scaling features, or handling outliers.
Step 3: Exploratory Data Analysis (EDA)
Now comes the fun part: diving into the data to uncover insights, patterns, and trends. EDA is an iterative process of asking questions, visualizing the data, and interpreting the results to gain a deeper understanding of the dataset and domain.
Some key questions we might seek to answer about the fitness tracker market include:
- Which brands have the most listings and market share?
- What is the distribution of prices across different brands and models?
- How do average ratings compare between the top brands?
- Is there a correlation between price and rating or number of reviews?
- What are the most common features and selling points of the best-selling trackers?
Let‘s answer these questions using Python‘s data manipulation and visualization libraries. Make sure you have Matplotlib and Seaborn installed:
pip install matplotlib seaborn
Import the required libraries:
import matplotlib.pyplot as plt
import seaborn as sns
Brand Market Share
First, let‘s see which brands dominate the fitness tracker market in terms of sheer number of listings.
brand_counts = df[‘brand‘].value_counts()
plt.figure(figsize=(10,5))
sns.barplot(x=brand_counts.index, y=brand_counts)
plt.title(‘Number of Listings by Brand‘)
plt.xlabel(‘Brand‘)
plt.ylabel(‘Number of Listings‘)
plt.xticks(rotation=45)
plt.show()
From this chart, we can see that Noise, boAt, and Fire-Boltt are the three brands with the most fitness tracker listings on Amazon India, suggesting they have the largest market share. Interestingly, many of the global market leaders like Apple, Fitbit, and Garmin have relatively few listings, likely due to their higher price points.
Price Distributions by Brand
Next, let‘s examine the distribution of prices for the top brands to get a sense of their positioning and target audiences.
top_brands = brand_counts.index[:5]
df_top = df[df[‘brand‘].isin(top_brands)]
plt.figure(figsize=(10,5))
sns.boxplot(x=‘brand‘, y=‘price‘, data=df_top)
plt.title(‘Price Distribution by Brand‘)
plt.xlabel(‘Brand‘)
plt.ylabel(‘Price (INR)‘)
plt.show()
This box plot reveals that Noise and boAt offer the most affordable options, with median prices around Rs. 1,500, while Fire-Boltt and Amazfit cater to slightly higher price segments. Fossil clearly targets the premium end of the market with its smartwatches.
Average Ratings Comparison
Ratings are a key factor influencing consumer purchase decisions. Let‘s compare the average ratings of the top brands.
brand_ratings = df_top.groupby(‘brand‘)[‘rating‘].mean()
plt.figure(figsize=(8,5))
sns.barplot(x=brand_ratings.index, y=brand_ratings)
plt.title(‘Average Rating by Brand‘)
plt.xlabel(‘Brand‘)
plt.ylabel(‘Average Rating‘)
plt.show()
We can see that despite its higher prices, Fossil actually has the lowest average rating of 3.9, while boAt and Noise lead the pack with ratings above 4.2. This suggests that consumers are highly satisfied with the value proposition of the more affordable brands.
Correlations between Price, Rating, and Reviews
Do more expensive fitness trackers necessarily garner better ratings or more reviews? Let‘s find out using scatter plots and correlation coefficients.
plt.figure(figsize=(8,5))
sns.scatterplot(x=‘price‘, y=‘rating‘, data=df_top)
plt.title(‘Price vs. Rating‘)
plt.xlabel(‘Price (INR)‘)
plt.ylabel(‘Rating‘)
plt.show()
print(df_top[‘price‘].corr(df_top[‘rating‘]))
The scatter plot and weak positive correlation of 0.24 suggest there isn‘t a strong relationship between price and rating for these top fitness tracker brands. More expensive doesn‘t necessarily mean better reviewed.
plt.figure(figsize=(8,5))
sns.scatterplot(x=‘price‘, y=‘reviews‘, data=df_top)
plt.title(‘Price vs. Number of Reviews‘)
plt.xlabel(‘Price (INR)‘)
plt.ylabel(‘Number of Reviews‘)
plt.show()
print(df_top[‘price‘].corr(df_top[‘reviews‘]))
Interestingly, we see a weak negative correlation of -0.31 between price and number of reviews, implying that cheaper fitness trackers tend to receive more reviews, likely due to their greater affordability and popularity.
Top-Selling Features
Finally, let‘s dive into the product details to surface the most common features and selling points of the best-reviewed trackers.
top_trackers = df.nlargest(10, ‘rating‘)
features = []
for desc in top_trackers[‘description‘]:
features.extend(desc.split(‘, ‘))
from collections import Counter
feature_counts = Counter(features)
top_features = [pair[0] for pair in feature_counts.most_common(5)]
feature_freq = [pair[1] for pair in feature_counts.most_common(5)]
plt.figure(figsize=(10,5))
plt.pie(feature_freq, labels=top_features, autopct=‘%.0f%%‘)
plt.title(‘Top 5 Features of Best-Rated Fitness Trackers‘)
plt.show()
Based on this analysis of product descriptions, heart rate monitoring, sleep tracking, blood oxygen monitoring, long battery life, and water resistance emerge as the most touted features of the top-rated fitness trackers.
Key Takeaways and Recommendations
Through this step-by-step EDA of the Indian fitness tracker market using Python, we‘ve uncovered several key insights:
- Indian brands like Noise, boAt, and Fire-Boltt dominate the budget to mid-range smartwatch segment, while global players focus on the premium end
- Consumers are highly satisfied with the affordable options from Noise and boAt, which boast ratings above 4.2
- Price doesn‘t necessarily correlate with higher ratings or more reviews; in fact, cheaper trackers tend to receive more reviews
- Heart rate monitoring, sleep tracking, SpO2 monitoring, long battery life, and water resistance are the key selling points for top-rated trackers
Based on these findings, we can offer a few recommendations for brands and retailers operating in this space:
- Prioritize affordability and value for money, as Indian consumers have embraced budget-friendly smartwatches
- Focus marketing and product development efforts on the core features of heart rate, sleep, blood oxygen, battery life, and water resistance
- Encourage customer ratings and reviews, especially for lower-priced products, as they heavily influence purchase decisions
- Consider partnering with Amazon and other leading e-commerce platforms, as that‘s where the majority of sales and reviews are happening
Of course, this is just scratching the surface of what‘s possible with data science and machine learning in the wearables domain. Brands could further leverage NLP techniques to mine customer reviews for sentiment and product feedback, build recommendation systems based on consumer preferences, forecast demand and optimize pricing using time series models, or even build apps to provide personalized health and fitness insights to consumers.
Aspiring data scientists and fitness enthusiasts can try web scraping and analyzing tracker data on their own by following the steps outlined in this guide. Experiment with different e-commerce sites, data cleaning techniques, visualizations, and statistical analyses to build your skills and discover new insights.
The global fitness tracker market may be crowded and competitive, but by harnessing the power of data, brands that adapt and innovate can capture the hearts (and wrists) of today‘s health-conscious consumers.
![Working with Missing Data in Python [Explained in 5 Steps]](https://33rdsquare.com/wp-content/uploads/2024/12/20241205224238-67522c5eb1fda.jpg)
