Keyword Cluster (KeyBERT) Looker Studio Demo Visualization Dashboard for SEO

This interactive Data Studio (formerly Looker Studio) dashboard turns raw keyword data into visual insights that reveal patterns, clusters and opportunities in your keyword research. Part of Lazarina Stoy’s MLforSEO course resources, it helps you make sense of hundreds or thousands of keywords by visualising cluster sizes, search volumes, intent and difficulty in ways that spreadsheet rows simply can’t convey.

The demo runs on a small sample of around 1,150 keywords that were labelled with KeyBERT: each keyword has a core 1-gram (for example “flotation” or “tank”) and a core 2-gram (such as “float therapy”) that act as its clusters.

What’s in the dashboard

• Filters: keyword, core 1-gram, core 2-gram, SERP features and primary intent, plus sliders for search volume, keyword difficulty and CPC.
• Treemaps: 1-word and 2-word clusters sized by number of keywords and by search volume, so high-value topic areas stand out at a glance.
• Sankey diagrams: the top 50 relationships between 1-word and 2-word clusters, by keyword count and by search volume.
• Intent breakdown: a stacked bar chart showing the commercial, informational, transactional and navigational split of search volume for the top 50 clusters.
• Metric exploration table: each keyword with its clusters, primary intent, SERP features, keyword difficulty and volume, colour-coded for quick pattern recognition.

How to use it with your own data

1. Run your keyword list through the KeyBERT for SEOs Colab notebook to get the Core (1-gram) and Core (2-gram) columns.
2. Add volume, keyword difficulty, CPC, primary intent and SERP features from your keyword research tool.
3. Load the file into Google Sheets, make a copy of the report and swap in your sheet as the data source, matching the field names.
4. Filter to a cluster, read its intent mix and difficulty, and decide whether it deserves its own page, a section or nothing at all.

Use this for

• Identifying natural keyword clusters and topic groups within large keyword datasets to inform content strategy
• Visualising search volume distributions to prioritise which topics deserve the most attention
• Presenting keyword research findings to stakeholders in a visual, easy-to-understand format
• Spotting high-volume, low-difficulty keyword opportunities that might get lost in a spreadsheet

Tips and limits

KeyBERT labels each keyword with its most representative word or phrase, so spelling variants become separate clusters: in the demo, “flotation” and “floatation” sit apart. Merge variants before you visualise. Some keywords get no 2-gram label and show as null. The Sankey and bar charts show the top 50 clusters only, so use the filters for the long tail.

This is perfect for SEO specialists and content strategists managing large keyword portfolios who want to communicate keyword insights visually to their teams or clients.

Learn it on MLforSEO

Read how to do keyword clustering with KeyBERT, then take the clustering module in Introduction to Machine Learning for SEO.