What data was used for analysis?

Experts highlight various types of data used for analysis across different fields:

1. Unstructured Data:

   • Text, images, video, documents, and emails comprise about 80% of collected data. Techniques like natural language processing, semantic analysis, and neural networks (recurrent and convolutional) are used to analyze this data after extraction processes 1.

2. Financial Data:

   • Data sets such as the FTSE 100 and DAX indexes were analyzed to obtain pricing data and calculate correlations. This process involved pre-processing data, using libraries like Gensim, and applying linear regression to determine the usefulness of signals from the data 2.

3. Computer Vision Datasets:

   • Evaluations of machine learning tools involved creating controlled datasets, like a dress dataset for Urban Outfitters, and using public benchmark datasets like CIFAR-10, MNIST, and Fashion-MNIST. This approach allowed for assessing performance and usability across different models 3.
     • 

       1

       Analyzing Unstructured Data

       00:09:13
     • 

       2

       Analyzing Financial Data

       00:09:43
     • 

       3

       Evaluating ML Tools

       00:26:14
       
     • 

       4

       Dataset Analysis Insights

       00:22:22
       

4. Qualitative Analysis:

   • A study involving 100 data sets and 14 highly cited data sets coded for 100 variables, examining aspects like licensing, privacy, and ethical considerations. Grounded theory was used to analyze themes in the data's construction, revealing biases and thematic focuses such as universality versus particularity, speed over care, and model building over data set preparation 4.