We Do Not Know Phone Number List
We do not know the approximate distribution of the samples used for cluster analysis, nor do we know which categories the system will classify them into, and there may not be any relevant category information for reference in advance Therefore, cluster analysis is more like a method for establishing hypotheses, and other statistical methods are needed to test related hypotheses In the process of generating user portraits, it is recommended to use cluster analysis as a way to explore the classification structure and provide data support rather than Phone number list (and may not) rely entirely on cluster analysis to form end-user classification conclusions
The following is a detailed description of the application of cluster analysis in user portraits based on a user classification case in the redesign of a financial lending service process 01 Applicable data types for cluster analysis The data types used in cluster analysis are mainly multi-dimensional, continuous/rank/categorical variables, and the data volume is required to be large enough and objectively measurable Therefore, it is more suitable for the situation where researchers already have massive, multi-dimensional user objective data Data sources include: product backend data that has been in operation <a title="Phone Number List" href=" removed link " target="_blank" rel="noopener">Phone Number List for a period of time , e-commerce browsing and purchasing behavior data, customer CRM data, WeChat public account backend data , etc Based on these data, we can classify users into Phone number list several categories based on objective data such as actual user behavior data (such as clicks, forwarding times, usage frequency, etc ), demographic data and other data Therefore, cluster analysis is widely used in research projects focusing on population, market and consumer behavior, such as consumer behavior research, market segmentation research, and e-commerce operation strategy research Cluster analysis involves the process of user classification:
<img src=" removed link " />
In user research work, user classification can be based on qualitative or quantitative data, but will eventually converge to a specific, clear, empirical classification model that can serve future products Design and Operation Figure 2: The application Phone number list of cluster analysis in user classification portraits——based on the application ideas and cases of psychostatistics Figure 2: Group portrait of the 2019 WeBank user survey Data alone cannot help us define and interpret sample profiles under different categories, nor can statistical results be directly applied to production design and operations Therefore, the method of cluster analysis should be carried out in combination with qualitative research (such as product walkthrough, Phone number list user interview, internal interview, observation, workshop, etc ) and quantitative research (questionnaire survey, interview survey, acceptance test, etc ) In this case, the researchers adopted a qualitative method, followed by clustering, and then supplemented quantitative methods to form and apply the results of cluster analysis, as shown in Figure 3: The application of cluster analysis in user classification portraits——based on the application ideas and cases of psychostatistics Figure 3: Financial Lending Service Process Redesign - User Portrait Creation Process
Thanks for commenting on our Analytica User Forum! Do you have a specific question? Or something we can help you with or weigh-in on? We just wanted to make sure we are properly assisting you, if you wanted assistance.
Let us know. Happy to help and happy modeling!