In this blog post, I'm going to compare some of the largest tech company's Instagram page's using my open source Python library instascrape! We'll be exploring their respective engagements, followers, amount of posts, etc. 🙌

The companies we'll be comparing for this exercise are

• Google
• Apple
• IBM
• Facebook
• Microsoft
• Adobe
• Oracle

Scraping the data

First, let's start by getting a list of their usernames:

companies = ["google", "apple", "ibm", "facebook", "microsoft", "adobe", "oracle"]

Now, scraping our data is as easy as

from instascrape import Profile 
profiles = [Profile(username) for username in companies]
for prof in profiles: 
    prof.scrape()

And that's it! We just scraped 364 data points from 7 profiles with just a few lines of code, let's use the to_dict method to get a list of dict's that can be passed into a pandas.DataFrame for expressive and powerful data analysis.

import pandas as pd 
data = [prof.to_dict() for prof in profiles]
df = pd.DataFrame(data)

Exploring our data

First, let's start by comparing how many followers each page has using a matplotlib bar plot:

import matplotlib.pyplot as plt 
plt.style.use("seaborn-darkgrid")
plt.bar(df["username"], df["followers"]) 

We can immediately see that Apple clearly has the most followers and surprisingly, Facebook doesn't have as many as one might expect.

Now let's see who has the most amount of posts:

import matplotlib.pyplot as plt 
plt.style.use("seaborn-darkgrid")
plt.bar(df["username"], df["posts"]) 

Finally, we're going to examine each page's engagement as a function of time and see how the different pages are doing

(NOTE: some of the specifics in the code are skipped so we can focus on what's important; additionally Apple will not be pictured as their data is significantly larger)

for prof in profiles:
    posts = prof.get_recent_posts()     #gets the 12 most recent posts
    posts_data = [post.to_dict() for post in posts]
    post_df = pd.DataFrame(posts_data)
    plt.plot(post_df.upload_date, post_df.likes, label=prof.username)

Some interesting things we can see right off the bat are:

• Oracle barely gets any likes
• Surprisingly neither does Facebook
• Adobe, Google, and Microsoft post relatively frequently
• IBM hasn't posted in almost two weeks
• Microsoft gets the most likes on average on their posts

Conclusion

And that's pretty much it! This is just a small taste of what instascrape can accomplish and it's up to you with how you use it so get out there and start exploring that data!

If you like what you read, check out some of my other posts 😄

Scraping 25,000 data points from Joe Biden's Instagram using instascrape

Chris Greening ・ Nov 5 '20

#showdev #python #datascience #contributorswanted

Downloading recent Instagram photos using instascrape and Python

Chris Greening ・ Oct 26 '20

#python #webscraping #showdev #contributorswanted

Also, check out the official repository and drop it a star ⭐ or contribute!