MongoDB's aggregation pipeline processes data through sequential stages. The $match stage filters documents similar to WHERE in SQL. $group groups documents for aggregation operations like $sum, $avg, $count. $sort orders documents by specified fields. $project reshapes documents with inclusion, exclusion, or computed fields. $lookup performs left outer joins with other collections. $unwind deconstructs arrays into separate documents. $addFields adds new fields to documents. $bucket categorizes documents into groups based on boundaries. $facet enables multi-faceted aggregation within a single stage. $replaceRoot replaces document with a sub-document. The pipeline supports geospatial operations with $geoNear. Use indexes on fields used in $match and $sort for performance. Optimize pipeline order: filter early with $match. The aggregation pipeline runs on the database server, reducing data transfer. Limit results with $limit stage. Skip results with $skip for pagination. Aggregation results can be output to a new collection with $out. MongoDB Compass provides visual aggregation builder. The aggregation pipeline is MongoDB's most powerful data tool.