We have large sets of qualitative data that need to be analysed. I was planning to use NVivo or similar, which I'm familiar with. However, I've been recommended Python...but I'm sceptical of the quality of an automated process for qualitative analysis. Does anyone have feedback on how it compares to more traditional methods of qualitative analysis? I'm sure it doesn't pick up the nuances...however, the analysis is of qualitative data (mostly community sentiment and staff notes) collected by a large NGO, which we want to analyse for needs assessment, programming improvement/development, and evaluative research...so the deep level of nuance I usually want probably isn't really as relevant...maybe! Thanks in advance.
Scikit-learn (Sklearn) is the commonly used library for ML in Python for predictive data analysis, providing ML and statistical modeling including classification, regression, clustering and dimensionality reduction via a consistence interface. Link: https://scikit-learn.org/stable/
You may follow http://makemeanalyst.com/data-science-with-python/python-libraries-for-data-analysis/ for somme additional libraries like Numpy and Scipy.
Hope this is useful.
Dear Leanne Kelly,
Python is a general and easy to read programming language. It is used in content analysis because it is simple and well documented. Plus, it's easy to learn to set up and use, and it's handy for anyone who wants to automate repetitive tasks. It can be intended for qualitative research, it is designed to facilitate the examination, analysis and understanding of large corpuses. It is therefore designed for the analysis of textual content.
For more information about this subject, i suggest you see links on the topic.
https://www.canada.ca/en/privy-council/services/public-engagement-resources/content-analysis-consultations/python-content-analysis.html
https://www.kaianalytics.com/post/how-to-use-text-analysis-techniques-bring-qualitative-data-to-life
https://www.reddit.com/r/learnpython/comments/a95skn/qualitative_data_analysis_and_pyqda/
https://www.futurelearn.com/info/courses/data-analytics-python-learning-python-functions/0/steps/190359
Best regards
The time it would take you learn Python and write such a complex program would undoubtedly be wasted. People who say that Python is "easy to learn" mean that it easy compared to other advanced programs such as C++. Given the amount of effort that has gone into developing high-grade programs such as NVivo, you would probably be reinventing the the wheel at best.
Instead, I suggest that you investigate the "big data" features within NVivo, which allow you do do a degree of automated coding and theme building.
Ah David L Morgan that's a great idea and sits much better with me. I'm certainly not a computer programmer by any stretch of the imagination so I'm not too keen on learning a new program. I am really curious to compare the outcomes of the two - would be a great paper seeing as there doesn't seem to be anything else like it that I can see from a quick look in Google Scholar. If any Python experts want to test the outcomes of NVivo versus Python for qualitative analysis, we could analyse the same sample of text and see what happens, and if it's interesting we could write it up! It would be really interesting to compare the differences in terms of time investment, ease of use, and depth of analysis. Thanks all for your responses too :)
NLTK (https://www.nltk.org/) is widely used toolkit for python programmers to do natural language processing (NLP) or textual data analysis.. however as you said, you are not from programming background nor you are interested to learn a new programming language therefore its best to use some well established tools meant for such analysis (NVivo)..
We can actually collaborate and try python or R language to conduct qualitative analysis on the data corpus that you have collected to compare various parameters as mentioned by you...
That would be amazing Shivsevak Negi ! I’d love to collaborate if you’re up for it. It would be great to compare different tech. Thank you!
Please suggest, how should we proceed from here.... as far as our collaboration is concerned?
Do you want to just send me an email to [email protected] so that I have your email Shivsevak Negi - We are just starting this work so it will take us a while to get NVivo sorted out and get up and running and then we can go from there. Looking forward to collaborating. Cheers, Leanne
Many programming languages, not only Python, can be used for qualitative data analysis. ... Commonly used Python libraries include instructions for techniques, such as natural language processing, sentiment analysis, and topic modeling, that help analyze content.
Let me share Stefanie Lai's experience with you. https://medium.com/@steflaihk/how-to-use-python-and-ai-in-qualitative-research-in-excel-96343f4dbdb6
- Badges
- Science topic
- Similar topics
- Filing