As a freelance AI consultant, I strive to optimize every aspect of my work, especially when it comes to my blog. When I started my blog at Evert’s Labs, I realized that SEO (Search Engine Optimization) was a major component in driving traffic and growing my audience. However, managing the SEO tasks manually was time-consuming. That’s when I decided to take things up a notch and develop a Python program to automate various SEO tasks, saving me both time and money.
In this blog post, I’ll take you through how I created a Python program that automates SEO for my blog. I’ll explain the components of this system, the code behind it, and how it has made a huge difference in efficiency and cost savings.
The Key Features of My SEO Automation System
- Autocategorization
Organizing content into the right categories is vital for SEO. Manual categorization takes up a lot of time, especially as the number of posts grows. I used a Python program to automate the categorization of blog posts based on their content. By analyzing keywords and context, the system assigns posts to appropriate categories automatically. - Autotagging
Tags help search engines understand the context of a blog post and improve search visibility. Instead of manually adding tags, my program scans the content and generates relevant tags. It uses natural language processing (NLP) techniques to identify the most important terms in the text and generates tags accordingly. - Autolabeling of Media Files
Images, videos, and other media files need to be properly labeled with alt text for SEO purposes. My automation program scans each media file and uses an image classification model to determine the best description for it. These labels are then applied automatically, ensuring SEO optimization without any manual input. - Autokeyword Optimization
Keyword optimization is one of the pillars of SEO. The program analyzes the blog post and suggests the best keywords to target, based on search volume and competition. It also helps in optimizing the placement of these keywords within the content, making sure that they appear in the right places (titles, headers, body text, etc.).
Code Breakdown and Explanation of API Calls
I used Python along with several libraries to achieve this automation. Here’s a breakdown of the core code and how the API calls work for each feature.
1. Autocategorization:
import openai
import pandas as pd
# Sample blog post content
content = "Exploring the latest AI technologies in healthcare..."
# API call to OpenAI for text categorization
openai.api_key = "YOUR_API_KEY"
response = openai.Completion.create(
engine="text-davinci-003",
prompt=f"Categorize the following content: {content}",
max_tokens=60
)
category = response['choices'][0]['text']
print(f"The category for this post is: {category}")This code uses OpenAI’s API to analyze the blog post content and suggest an appropriate category. The categorization is done by providing the content as a prompt to the AI model.
2. Autotagging:
from transformers import pipeline
# Initialize HuggingFace's BERT-based model for text tagging
tagging_pipeline = pipeline("zero-shot-classification", model="facebook/bart-large-mnli")
# Define possible tags
candidate_labels = ["AI", "Technology", "Healthcare", "Business"]
# API call to classify tags based on content
result = tagging_pipeline(content, candidate_labels=candidate_labels)
tags = result['labels']
print(f"Suggested tags: {tags}")This uses a pre-trained BART model from HuggingFace to suggest relevant tags for the blog post based on the content. The model classifies the content into predefined tags and outputs the most relevant ones.
3. Autolabeling of Media Files:
import torch
from torchvision import models, transforms
from PIL import Image
# Load pre-trained model for image classification
model = models.resnet50(pretrained=True)
model.eval()
# Load and preprocess image
img = Image.open("sample_image.jpg")
preprocess = transforms.Compose([
transforms.Resize(256),
transforms.CenterCrop(224),
transforms.ToTensor(),
transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
])
img_tensor = preprocess(img).unsqueeze(0)
# Predict image label
with torch.no_grad():
outputs = model(img_tensor)
_, predicted = torch.max(outputs, 1)
label = predicted.item()
print(f"Media file label: {label}")For autolabeling media files, I use a ResNet-50 model to classify images. The image is preprocessed and passed through the model to predict a label, which is then applied as alt text for SEO purposes.
4. Autokeyword Optimization:
from sklearn.feature_extraction.text import CountVectorizer
# Sample text for keyword optimization
content = "AI technologies are revolutionizing healthcare by enabling more efficient diagnostics..."
# Extract keywords using CountVectorizer
vectorizer = CountVectorizer(stop_words="english", max_features=5)
X = vectorizer.fit_transform([content])
keywords = vectorizer.get_feature_names_out()
print(f"Optimized keywords: {keywords}")This snippet uses scikit-learn’s CountVectorizer to extract the most important keywords from the content. These keywords are used to optimize the blog post further.
Time Savings and Efficiency
By automating these SEO tasks, I have saved several months of manual work. Before automation, I spent hours categorizing posts, adding tags, labeling media, and optimizing keywords. Now, all of these tasks are done in minutes, allowing me to focus on more important aspects of my work, like content creation and AI development.
Cost Savings with Ollama and Llama 3.5
One of the significant benefits of this automation system has been cost savings. Initially, I used OpenAI and Claude for API calls, but the costs quickly added up. I then switched to using Ollama, which leverages Llama 3.5, a more affordable alternative to OpenAI and Claude. By using Ollama, I was able to cut down on my API expenses without sacrificing the quality of the AI’s output.
The performance of Llama 3.5, combined with Ollama’s user-friendly interface, allows me to perform the same tasks at a fraction of the cost. The difference in pricing has been significant, providing excellent value for my SEO automation needs.
Conclusion
Automating SEO for my blog using Python has been a game changer. It has saved me countless hours while improving the SEO quality of my posts. The ability to categorize, tag, label, and optimize content automatically has taken my blog to the next level, and the cost savings with Ollama and Llama 3.5 have been substantial. This project has shown me that with the right AI tools and a little bit of Python magic, anything is possible.
I, Evert-Jan Wagenaar, resident of the Philippines, have a warm heart for the country. The same applies to Artificial Intelligence (AI), machine learning and data analysis. I have extensive knowledge and the necessary skills to make the combination a great success. I offer myself as an external advisor to the government of the Philippines. Please contact me using the Contact form or email me directly at evert.wagenaar@gmail.com!