Depop App Review Analysis

How I Turned App Reviews Into My Golden Ticket to Depop

Have you ever had one of those "Aha!" moments where a lightbulb practically explodes above your head? That was me during the final round of my Depop interview. 💡

Picture this: I had two days to prep a project that could wow a room full of senior data scientists. The problem? I wasn’t feeling it with my existing projects. I knew I needed something different—something fresh, bold, and completely Depop-worthy.

And then it hit me. Why not make my project all about Depop itself?

The Plan: A Data Deep-Dive

Armed with Python and my trusted data analysis toolkit, I dove head first into the world of app reviews. I scraped 19,303 reviews (yes, nineteen thousand), spanning a decade of user feedback from the Google Play Store. Here's a running bar chart of the app ratings over the years.

(Psst .. Wait for a bit for the bar chart race to load, please🙃)

It was a treasure trove of insights—compliments, complaints, and downright savage critiques. My mission was to decode this chaos to uncover why the app’s rating had dropped and how it could bounce back.

Data, Data Everywhere

The reviews were messy—like, “Who even writes this stuff?” messy. Here’s how I tackled it:

• Step 1: Removed unnecessary columns. Bye-bye clutter. 👋

• Step 2: Cleaned up the text. Lowercase, punctuation gone, stop words eliminated.

• Step 3: Tokenized the words for analysis. Think of this like turning a tangled spaghetti of text into neat little meatballs of meaning.

  
  

The Fun Part: Experimentation

Here’s where things got juicy. I asked myself three big questions:

1. Why did the app’s rating dip?

2. What’s bugging users the most?

3. How can Depop reclaim its glory days?

   
   

I used some fancy tools—LDA, RoBERTa, and NRC sentiment analysis—to break down reviews into themes. Let me tell you, it was like peeling an onion (and yes, there were tears involved).

What Users Loved:

• “Love this app, so easy to use!”

• “Great for alternative fashion. Customer service is amazing!”

  
  

What Users Hated:

• Scammers: “The first person I bought from was a scammer. It took months to get my money back!”

• Technical glitches: “Not getting the 6-digit code to log in. SO FRUSTRATING!”

• Hidden costs: “It says Depop takes 10%, but half my earnings vanish. Why?”

  
  

Key Findings (and My Two Cents)

• User Interface: People loved the app’s simplicity but wanted better sorting options.

• Tech Issues: Frequent login problems? Not a good look.

• Scammers: They’re everywhere, and they’re annoying everyone.

• Costs: Hidden fees? Not cool.

  
  

With all these insights in hand, I crafted actionable recommendations, from refining payment features to cracking down on scammers.

The Results: A Mic-Drop Moment

When I presented my findings, I could see the wheels turning in the senior data scientists’ heads. They got it. They saw how these insights could genuinely help Depop improve.

Fast-forward to today, and guess what? The app’s rating climbed from 3.8 to 4.3. I’m not saying my project single-handedly saved the day, but I’d like to think it was a pretty solid nudge in the right direction. 😏

Takeaway: The Power of a Personal Touch

This project taught me two big things:

1. Data doesn’t lie—it’s a goldmine of insights if you’re willing to dig deep.

2. When in doubt, make it personal. My twist on the project wasn’t just about showing off skills; it was about showing I cared.

   
   

This was more than a portfolio project—it was my golden ticket to Depop. And honestly? It was worth every late-night hour and every line of code.

Project Github Link