CONTACE INFORMATION:
- Aliyun mail: jidong@zhgjaqreport.com
- Blog: https://zhgjaqreport.com
- Gmail：Jidong694643@gmail.com
- Proton mail：Jidong694643@proton.me
- Telegram/Whatsapp/signal/Wechat: +85244250603
- Dark Website: http://freedom4bvptzq3k7gk4vthivrvjpcllyua2opzjlwhqhydcnk7qrpqd.onion
The triad for OSINT in China encompasses collection, analysis, and dissemination. More than 2.5 exabytes of data is collected every day from social media, satellite images, etc.; NLP algorithms do sentiment analysis for 2 million texts daily, and encrypted API interfaces achieve real-time intelligence pushes in less than 10 minutes.
Collection
Information gathering is facing enormous quantities of data and multi-dimensional information channels. In 2024, the whole world will create about 2.5 exabytes per day, and Chinese Internet users contribute 28% to the whole world’s traffic. The number of articles updated every day on Weibo and WeChat public accounts alone is more than 50 million, while the blowing content of social platforms is distributed in an extremely wide manner. Precisely screen these open sources and ensure data quality.
A 2023 social data study showed that the peak of public opinion on hot events usually reached its peak within 48 hours. After certain emergencies, such as the release of public health announcements or new policy implementation details in a certain place, the keyword search volume on social platforms can surge 500% within 1 hour. Because of this, the use of data crawling scripts and an automated system with simple keywords is indispensable in being efficient. Python and R are usually applied to perform these tasks, especially for crawling in the API interface of social platforms, which have responses with millisecond-level speed.
From the high-resolution satellite images released by Global Observation Alliance, the latest 0.5 meters/pixel detail accuracy has shown certain vehicle trajectories and outlines of new buildings after analyzing the dynamic changes in airports, ports, and military facilities. According to incomplete statistics, every year, local governments in China publish more than 20,000 kinds of yearbooks and development reports covering economic indicators, social development statistics, construction plans, and other information. While these kinds of documents cannot be acquired simply by web crawling, they often provide more realistic and precise historical data for policy interpretation and economic trend forecasting.
During the COVID-19 pandemic, there were multiple versions of data on case growth on social platforms, with one time the difference between the number of confirmed cases announced by the official and private statistics reaching as high as 15%-20%. It needs multiple cross-validation before analysis, especially for information of unknown or ambiguous data sources; best avoid those as much as possible. After two or more multi-source comparisons, an error rate reduction of 32% can be achieved.
Analysis
The analysis phase in OSINT involves structured processing and multi-dimensional comparison of data. According to industry statistics in 2023, about 67% of the errors in intelligence come from the analysis phase, not from inaccurate data itself.
The general method for the data analysis industry is to use regression models, clustering algorithms, text analysis systems, and so on, to turn unstructured data into quantifiable indicators. In order to assess the effectiveness of policy implementation in a certain region, through trend analysis of employment rate data for 12 consecutive months, it was found that in the third month after the release of a policy, the employment rate rose by an average of 2.8%, but began to fall back in the sixth month, showing that the obvious initial effects of the policy lacked long-term support.
Social media: more than 4.5 billion tweets and comments are generated every day. To do this efficiently, intelligence analysts typically use NLP algorithms to conduct sentiment analysis and content classification on large sample data. According to an online public opinion analysis on a hot topic, out of 2 million comments about the topic, 78% of the content is focused on speculation related to the aftermath of the incident; only 10% involve hard facts.
Data such as the number of new cases per day and virus transmission rate can be modeled in epidemic prevention and control to predict the growth trend in the coming week. In the year 2020, based on medical data of a certain region, the ARIMA forecasting model was established to predict that in the next 7 days, the peak cases would reach 1,200, while in fact, there was only a difference of 3.5% between the actual data and the forecast.
It is on the core linkage of distinguishing data trends from traps by doing the correlation and casual analysis. For example, it was noted while analyzing the corporate financial data that during a quarter the revenue was increased by 15%, whereas the marketing expenses for that period have increased by 30% with an increase of merely 5% in market share. Such asymmetric growth triggered reflections among the analysts with questions regarding efficiency about the market launch strategy. It was finally found that though more than 90% of the target customer groups were covered by the invested advertising channels, the conversion rate was below the industry average.
Dissemination
In OSINT, information dissemination will determine whether obtained intelligence can be timely and effectively grasped and utilized by the related departments, institutions, or individuals. Based on the 2023 Global Intelligence Industry Report, enhancing efficiency in intelligence dissemination will shorten decision-making by 25% to 30%.
Real-time intelligence updates for those industries with very high demand-for instance, emergency management and financial markets-should be done hourly or even within 10-minute intervals for the proper capturing of the risk signals and for the proper measure to be taken. In the case of a large-scale cyber attack, sending real-time intelligence to the customer’s network monitoring platform in JSON format allowed it to achieve risk reminders on the whole network within 2 minutes, greatly enhancing response efficiency in contrast to a 10-minute delay due to the traditional release of a manual report.
Through synchronous dissemination across multiple platforms, such as email, instant messaging software, API interface integration, etc., it can be ensured that different departments obtain consistent information within a time window. In the case of a logistics company facing delays in global supply chains, Slack, Teams, and email systems were integrated for information synchronization. This helped reduce communication delays by 48% and shortened the average decision time from 72 hours to 24 hours.
Graphical reports are easier to understand than simple plain text reports, and it gives a speed increase in processing by about 40%. Now, if a think tank publishes an annual macroeconomic forecast, the interactive dashboard can show the change in trends for various indicators, thus enabling the user to adjust different parameters to get real updates of forecast model results in real-time.
Where transmitting sensitive information, one should apply a strict encryption protocol and an access control procedure. More than 70% of intelligence leaks occur because of the improper management of internal authority or unencrypted data transmission. The risks can be considerably reduced and the security of communication improved by applying end-to-end encryption, two-factor authentication, and auditing of data.
CONTACE INFORMATION:
- Aliyun mail: jidong@zhgjaqreport.com
- Blog: https://zhgjaqreport.com
- Gmail：Jidong694643@gmail.com
- Proton mail：Jidong694643@proton.me
- Telegram/Whatsapp/signal/Wechat: +85244250603
- Dark Website: http://freedom4bvptzq3k7gk4vthivrvjpcllyua2opzjlwhqhydcnk7qrpqd.onion