Context of data should be clear before data analysis

Before data analysis is carried out, it is important to understand the context of the available data. For example, if we analyze hotel reviews data and draw conclusions without understanding the context of the hotel reviews, then we could be drawing totally wrong conclusions. 

Let's assume that we have hotel reviews data. Firstly, we should ask several questions about the hotel reviews data before we try to find answers about business from hotel reviews data. Questions such as:

  • Who is the provider of the reviews? 
  • How are the reviews provided?
  • When are the reviews provided? 
  • What are the limitations involved in the review collection process?
  • In how many languages can guests provide reviews in the reviews system?
  • What % of the hotel guests provide reviews?
  • What % of the hotel guests were able to read at least one of the languages in which reviews collection system works?
  • Which channels are there for guests to provide reviews?
  • When did review collection start?
  • What is the sequence and timing of opening of the different channels of reviews?
  • What changes have happened in the review collection process?
  • How many unique hotel guests have provided reviews?
  • Are there reviews by third parties or competitors?
  • Are there any reviews that seem malicious?

This is not a comprehensive list of questions but sample of questions. Based on the answers to these questions, more refined questions should be asked about the data.  

Image by Mote Oo Education from Pixabay

Comments

  1. Another example I noticed while purchasing a book for my kid in amazon site. Someone had given a rating of 1 for R K Narayan's Malgudi days book. Only when I checked the comment, it was clear that the person has rated 1 out of 5 because of the slightly damaged condition of the book (due to delivery) and not because of the content of the book. If we only look at the ratings without understanding the context, surely we will draw wrong conclusions.

    ReplyDelete

Post a Comment

Thanks for your comment. It will be posted after checks.

Popular posts from this blog

ETL developer vs Data engineer

Mistakes and Improvements - Business Intelligence Demystified Book

Businesses should classify data not IT