Step 1 of 7
Move to the next step with this
Keyword Embedding Generation
Generate semantic embeddings for your keywords
info
Tip:Click the slideshow Slide icon above to see the educational content about this step!
What Are Embeddings?
Embeddings are mathematical representations of data that preserve relationships within your data.
Word → Embedding → Position
man
living being
feline
human
gender
royalty
verb
…
…
…
…
…
…
2D projection
x: royalty, y: inverse gender
man
woman
king
queen
codeBackend Code Structure
account_treeProcessing Pipeline
INPUT
keywords.csv
save_upload()
arrow_downward
load_and_clean_data()
→ Remove duplicates, validate
arrow_downward
generate_embeddings_parallel()
→ Split into batches
cloudOpenAI Batch 1
cloudOpenAI Batch 2
cloudOpenAI Batch 3
arrow_downward
write_output_csv()
→ Serialize vectors
OUTPUT
embeddings.csv
Example Data Transformation:
Input:
{
"Keyword": "best running shoes"
}
"Keyword": "best running shoes"
}
Output:
{
"Keyword": "best running shoes",
"Embedding": "[0.023, -0.15, ...]",
"Is_Brand": false
}
"Keyword": "best running shoes",
"Embedding": "[0.023, -0.15, ...]",
"Is_Brand": false
}
starsKey Features
speed
Parallel Processing
ThreadPoolExecutor for concurrent API calls
storage
File Persistence
Timestamped storage for inputs/outputs
verified
Data Validation
CSV structure validation and cleaning
auto_awesome
OpenAI Integration
text-embedding-3-small model