Brain Wikipedia Page Views for US Stocks

The dataset Brain Wikipedia Page Views has the objective to monitor the number of views and to measure the related attention on the Wikipedia pages views for the top largest 1000 US companies approximately corresponding to the Russell 1000 index. The raw number of views is transformed using a metric of called buzz to detect if a company is receiving more visits than average on various time horizons, e.g. the past day, week or month. The goal is to provide an alternative way to measure the attention of investors toward a specific company; this is complementary to the attention metrics measured from news or other sources (see for example the Brain Sentiment Indicator measuring buzz from financial news). The calculation steps are the following: - The Wikipedia pages for each company are mapped to the corresponding quoted entity with periodical updates. - Each company is mapped to multiple Wikipedia pages titled with variations of the company name in order to manage “redirects”; this is important to perform a comprehensive calculation of the page views. - On a daily basis the page views are aggregated at company level for various time intervals (e.g. the past day, week or month). The number of views for each company (similarly to the case of news) is typically proportional to the company size therefore a normalized measure is needed to measure the attention. - The attention on the company by investors or buzz is then calculated as a variations of page views with respect to the past average. The buzz is provided for various time intervals (past day, week or month).

搜集汇总

背景与挑战

背景概述

该数据集旨在通过追踪约1000家美国大型公司维基百科页面浏览量，量化投资者关注度。它采用标准化方法将原始浏览量转化为'Buzz'指标，以衡量不同时间周期内相对于历史平均的访问量变化，从而作为新闻情绪分析等传统关注度测量方式的补充工具。