Utilizing the natural langauage toolkit for keyword research
Download as PPTX, PDF3 likes1,204 views
This document discusses using the Natural Language Toolkit (NLTK) for keyword research and analysis. It provides instructions on installing NLTK and other Python libraries, preparing keyword data, and running scripts to classify and cluster keywords to identify trends and topics. The document demonstrates how to automate aspects of keyword research using NLTK to help analyze large datasets.
Utilizing the natural langauage toolkit for keyword research
- 1. Answer, why this presentation matters? Utilizing The Natural Language Toolkit for Keyword Classification & Cluster Analysis Miracle Inameti-Archibong | Erudite SLIDESHARE.NET/USERNAME @mira_inam
- 2. How to do keyword research and keep the will to live
- 3. B R I G T H O N S E O S U M M E R 2 0 2 1 Why Keyword Research.
- 4. Keyword research is the process of identifying which terms or phrases your target audience use to search for your goods, services.
- 5. Why do we need it?
- 6. Enables you analyse user behaviour in order to evaluate if your content satisfies the searcher’s intent.
- 7. Discover new Opportunity. Related products. Should filtered pages be optimised and indexable. How do your consumers' research.
- 8. Create a search and user- friendly information architecture.
- 9. What do we need for keyword research?
- 10. Keywords and search volume *other keyword Research Tools exist https://erudite.agency/insights/how-to-conduct-an-in-depth-
- 11. Analysing the data. Categories. Use case. Intent. @mira_inam #BrightonSEO
- 12. Manual keyword data sorting . Using excel, look for terms or phrases that occur frequently and tag them up in the desired category.
- 14. How to do keyword research and keep the will to live
- 15. B R I G H T O N S E O S U M M E R 2 0 2 1 Automating frequency tagging with the Natural Language Tool Kit (NLTK).
- 16. B R I G H T O N S E O S U M M E R 2 0 2 1 NLTK is a leading platform for building Python programs to work with human language data. Source - https://www.nltk.org/
- 17. Natural language processor with python Classification Tokenization Stemming Tagging Parsing Semantic Reasoning
- 18. Natural language processor with python Classification Tokenization Pull out various tokens from text and classify them by nouns, verbs
- 19. Natural language processor with python Instruction Determine categories based on word frequency
- 20. Everyday applications of NLP Chat boxes Voice commands on various devices Google’s knowledge graph – evolved from keyword lookup to language parsing
- 21. B R I G H T O N S E O S U M M E R 2 0 2 1 Getting started.
- 22. Download and install python library. Python.org/downloads/ (preferably version 3.6.4)
- 23. B R I G H T O N S E O S U M M E R 2 0 2 1 You also need the following python libraries.
- 24. Powerful data structures for data analysis, time series, and statistics. pandas 1.2.4 https://pypi.org/project/pandas/
- 25. A simple API for common natural language processing tasks such as part-of-speech tagging, noun phrase extraction, sentiment analysis, classification, translation, and more. https://textblob.readthedocs.io/en/dev/ TextBlob
- 26. Both libraries for reading data and formatting information from Excel files .xls format https://xlrd.readthedocs.io/en/latest/ https://openpyxl.readthedocs.io/en/stable/ &
- 27. B R I G H T O N S E O S U M M E R 2 0 2 1 How to pip install the python libraries
- 28. Pip install these libraries. Download requirement(1).txt file and save it on your computer (note the file path)
- 29. Open the command prompt. Open Windows Menu Type command prompt ❑In the Command Prompt Copy and paste - pip install -r C:somepathtorequirement s.txt @mira_inam #BrightonSEO
- 30. Paste the instruction and click enter. pip install -r C:somepathtorequi rementsrequirements .txt Highlighted in pink is the path where your file is saved @mira_inam #BrightonSEO
- 31. The analysis script. Download and save the python script
- 32. Prepare your keywords list. ❑Ensure your keywords list is clean, has no spaces or empty rows ❑Heading must be keywords ❑File must be .xlsx not CSV
- 33. Running the analysis. ❑Open a command prompt ❑Run the script ⮚ Python C:somepathtokeywords.py C:somepathtokeywordlist.xlsx C:somepathtotest_output.xls
- 34. Replace highlight with where your file is saved. C:somepathtokeywords.py- C:somepathtokeywordlist.xlsx
- 35. Replace highlight with where you want your test output file to be saved. C:somepathtotest_output.xls @mir_inam #BrightonSEO
- 36. Run the script. python C:UsersmiracDownloadskeywords.py C:UsersmiracDesktoptest.xlsx C:UsersmiracDownloadstest_output.xls
- 37. Output.
- 38. Sort the data to identify trends.
- 39. Spot topics & themes from the data.
- 40. B R I G H T O N S E O S U M M E R 2 0 2 1 Running the script on Google Colab
- 41. Make a copy and save on your google drive http://bit.ly/keywordclassification
- 42. Make a copy and save on your google drive Click on the folder to open upload data prompt
- 43. Upload your keywords data. Upload the file you want to run. Same rules about uploading a clean file still apply see slide 31
- 44. Very important – Keep all naming conventions the same. Keywords list – test.xlsx
- 45. Click run all or Ctrl+F9. Run all
- 46. Download the output file. Right click the 3 dots and download the output file
- 47. Download the output file. It can be a bit buggy so if you don’t see your output file wait a few minutes or re-run.
- 48. B R I G H T O N S E O S U M M E R 2 0 2 1 Cluster Classification
- 49. Pulls out the most common clusters by frequency in a histogram. Coffee Dresses
- 50. Pulls out word clouds.
- 51. Make a copy and save on your google drive http://bit.ly/keywordcluster How to use.
- 52. Running the data. Same way to upload data and run as slide 32-35
- 53. Increase the number of clusters depending on the size of the data. Click on cluster analysis to view the code.
- 54. Increase the number of clusters depending on the size of the data. Increase or decrease the number of clusters
- 55. No download. Scroll to the bottom of the script to use images of the cluster.
- 56. Feedback. What additional functions would you like to see? Send me a message @mira_inam .
- 57. Thanks to Eniola Olaleye @galileoeni Mahdi Mairza
- 58. View my slides here in full. erudite.agency/brighton-seo
- 59. Erudite Agency Ltd 2010-2020 © contact@erudite.agency Let’s start your growth journey. If you would prefer our help prioritising and consulting on your audit, or for more about how we can help you increase your visibility and exposure please e- mail or call: +44 (0) 1256 384890