In the field of Amazon product portraits, in addition to distribution analysis of reviews, prices, and rankings, you can also perform word frequency analysis on product listing titles.
What is word frequency analysis? Word frequency analysis is the statistical analysis of the number of times important words appear in text data, and is an important means of text mining. Its basic principle is to determine hot spots and their changing trends by the changes in the frequency of word appearance.
On the Amazon platform, the objects of word frequency analysis can be text content such as review text product titles, listing five-point descriptions, A+ graphic content, etc. Among them, word frequency analysis of product titles is of great significance in the field of product portraits.
In the Excel file of “Product Portrait”, product titles are listed as title texts of products with different search rankings.
In order to facilitate customized word frequency analysis, the 19,152 listing crawling results in the Excel file of “Product Portrait” are split into 192 separate Excel tables and saved in a folder named “Frequencies of Keywords”.
Different table files in the folder record the frequency analysis results of different ranking totals. For example, “Frequencies Of Keywords 100” indicates the frequency analysis results of the titles of the top 100 product listings in the search ranking; “Frequencies Of Keywords 200” indicates the frequency analysis results of the titles of the top 200 product listings in the search ranking; “Frequencies Of Keywords 1000” indicates the frequency analysis results of the titles of the top 1000 product listings in the search ranking; “Frequencies Of Keywords 19 152” indicates the frequency analysis results of the titles of the top 19152 products (all products) in the search ranking.
Based on operational experience, the frequency analysis results of product titles with different search ranking totals must be different. Therefore, a static comparative analysis is performed on the frequency analysis results of the products in the top 100 search rankings and the top 19152 products (all products) in the search rankings.
Open the Excel table “Frequencies Of Keywords 100” to see the frequency analysis results of the products in the top 100 search rankings.
For the comparative analysis of other words, it is not enough to rely on the results of two static word frequency analyses. Therefore, it is necessary to compare the word frequency analysis results of multiple search rankings to obtain a conclusion. To this end, the author uses Python to make two dynamic word frequency analysis arrangement charts in positive and reverse order for your reference. The video names are “Positive Dynamic Arrangement Chart” and “Reverse Dynamic Arrangement Chart”.
The “Positive Dynamic Arrangement Chart” shows the changes in the frequency of different title words when the search ranking is from small to large (from 100 to 19 152). Generally speaking, there are two meanings for new words/words with rapid ranking increases in the TOP 20 words: one is that medium and long-tail sellers tend to use such words; the other is that newly listed products generally have the selling points represented by the words. Therefore, operators can judge the differences in operating habits between medium and long-tail sellers and head sellers based on the “Positive Dynamic Arrangement Chart” video, and at the same time, they can judge the future trend of the category market based on the ranking changes of different selling point words.
The “reverse dynamic arrangement chart” shows the change in the frequency of different title words when the search ranking is from large to small (from 19,152 to 100). The changing trend of the words in the “reverse dynamic arrangement chart” represents the distribution pattern of the hot-selling product title text. If the operator wants to learn the skills of the top sellers in editing listing titles, he can optimize his own product titles by observing the changes of different words with the search ranking.
In addition to the above word frequency analysis methods, operators can also use the “word cloud chart” for static analysis. “Word cloud” was first used by Rich Gordon, associate professor of journalism and director of the new media major at Northwestern University in the United States in 2006. “Word cloud” is to visually highlight the “keywords” that appear frequently in network texts by forming a “keyword cloud layer” or “keyword rendering”.
If the “word cloud chart” is generated by combining the text of all product titles in the “product portrait” Excel file, the “word cloud chart” is obtained.
The more prominent (larger font size) words are in the “word cloud”, the higher their frequency of occurrence. For example, women’s, dress, sleeve, long, party, sleeveless and other words are high-frequency words, which is consistent with the above word frequency analysis results. Therefore, the “word cloud” is a visual display method of word frequency analysis, and its analysis results are no different from those of other word frequency analyses.