top of page

Text Data Analysis Using Microsoft Azure | Data Warehousing and Big Data Assignment Help



The assignment focuses on Big Data analytics on unstructured text data using Microsoft Azure. You are required to derive insights by applying big data distributed processing and machine learning techniques.


Data Set Details

The dataset contains ~10000 reviews of hotels in Vietnam. The fields are;


Field name Description

PostId Unique ID for each review

Subject The heading of the review

Rating The rating given by the user, ranging from 1 to 5

Hotel Name of the hotel

Hotel-Star Star rating of the hotel

Review Textual description of the review. The review field text has been pre-processed by removing non-alphanumeric characters and reduced the size to be max of 1000 characters.

Sentiment Sentiment value extracted for the ‘review’ using text analytics API in Microsoft

Cognitive Services. Sentiment score ranges from 0 to 1, 0 being negative and 1

being positive. (You may verify the sentiment score with the subject and review

fields.)

Sentiment Category Sentiment class generated based on the sentiment value.

< 0.9: Very positive

< 0.75 and <= 0.9: Positive

<0.5 and <= 0.75: Neutral

<= 0.5: Negative


What you are required to do?


1. HDInsight to aggregate reviews

Develop an aggregate of these reviews using your knowledge of Hadoop and MapReduce in Microsoft HDInsight.


a) Follow the same approach as the Big Data analytics workshop (using wordcount method in HDInsight) to determine the contributory words for each level of rating and sentiment category.


b) Present the workflow of using HDInsight (you may use screen captures) along with a summary of findings for each level of rating and sentiment category. MapReduce documentation for HDInsight is available here.


2. Azure Machine Learning for sentiment analysis

Using Azure ML Studio to cluster user reviews based on sentiment score. For text clustering, you should use the ‘review’ field. In the Filter-based feature selection module, use the ‘sentiment’ field in order to cluster reviews based on sentiment score. Download the cluster outputs into a CSV file to interpret the results and derive insights. You will need to calibrate algorithmic parameters by using different Numbers of Centroids and Distance Metric to derive meaningful clusters. Exclude sentiment, rating, or posted as selected columns to train the clustering model. Use only the preprocessed hashing features.



Provide the following,

a) A screen capture of the completed model diagram.

b) Details of parameters used for 1) feature hashing module, 2) filter based feature selection module and

3) K-Means clustering module

c) Details of the approach you chose for clustering and interpretation of clusters.


3. Findings

Summarise your findings from 1) and 2), on user rating, hotel rating, and sentiment towards accommodation options in Vietnam. Consider the challenges you faced in conducting Big Data analytics on a real-life text dataset.


Deliverables

1. A report on the three activities.

  • The report should be compiled in Microsoft Word only, font size 11.

  • Report should not exceed 10 pages. Diagrams, tables and any other visualizations/ screen captures should be in the main body of the report.

  • Make realistic assumptions on missing information and state these in the report.

2. A compressed folder of data files that would be useful to assess your work.



Codersarts is a top-rated website for students which is looking for online Programming Assignment Help, Homework Help, Coursework Help in C, C++, Java, Python, Database, Data structure, Algorithms, Final year project, Android, Web, C sharp, ASP NET to students at all levels whether it is school, college and university level Coursework Help or Real-time project.



If you have project or assignment files, You can send at contact@codersarts.com directly


Comments


bottom of page