MongoDB -- Portfolio Projects
Project 1: Blog Platform with Aggregation Pipeline
Level: Beginner | Time: 2 days | GitHub: mongodb-blog-api
Blog platform demonstrating MongoDB document design and aggregation pipelines.
javascript
// Schema design -- embed vs reference decision
// EMBED (comments in posts): reads are fast, updates are in one doc
// REFERENCE (authors as separate collection): author data shared, not duplicated
db.posts.insertOne({
title: "MongoDB Best Practices",
slug: "mongodb-best-practices",
author_id: ObjectId("..."),
tags: ["mongodb", "nosql", "database"],
content: "...",
comments: [ // Embedded -- only 50 comments expected max
{user: "alice", text: "Great post!", created_at: new Date()}
],
stats: {views: 0, likes: 0},
created_at: new Date()
});
// Aggregation: top authors by total views
db.posts.aggregate([
{ $group: {
_id: "$author_id",
total_views: { $sum: "$stats.views" },
post_count: { $count: {} },
avg_views: { $avg: "$stats.views" }
}},
{ $sort: { total_views: -1 } },
{ $limit: 10 },
{ $lookup: {
from: "authors",
localField: "_id",
foreignField: "_id",
as: "author"
}}
]);
Steps: Schema design decisions documented, aggregation pipeline, text search index, Atlas free tier
Project 2: Real-Time Analytics with Change Streams
Level: Intermediate | Time: 2-3 days | GitHub: mongodb-change-streams
Listen to MongoDB changes in real-time and react -- no polling, instant updates.
python
import pymongo, threading
client = pymongo.MongoClient("mongodb://localhost:27017")
db = client.myapp
def watch_orders():
pipeline = [
{"$match": {"operationType": "insert",
"fullDocument.status": "paid"}}
]
with db.orders.watch(pipeline) as stream:
for change in stream:
order = change["fullDocument"]
print(f"New paid order: {order["_id"]} total: {order["total"]}")
send_confirmation_email(order)
update_inventory(order["items"])
thread = threading.Thread(target=watch_orders, daemon=True)
thread.start()
Steps: Change streams on orders collection, update downstream systems, handle resume tokens for restart
Portfolio Checklist
•[ ] Document design decisions explained in README
•[ ] Indexes for every query pattern
•[ ] Aggregation pipeline with at least 5 stages
•[ ] Change streams or Atlas triggers for real-time processing
