Developing Social Media in Ukraine: Statistical Forecasting

• Main Author: O. TUMANOV
• Format: Article
• Language: Ukrainian
• Published: National Academy of Statistics, Accounting and Audit 2020-03-01
• Series: Naukovij Vìsnik Nacìonalʹnoï Akademìï Statistiki, Oblìku ta Auditu
• Subjects: forecasting methods, analysis, social media, holt method, facebook.
• Online Access: https://nasoa-journal.com.ua/index.php/journal/article/view/203

Predictions and divination have always interested people. Since ancient times, oracles and wizards have answered questions that interest people in completely different areas of life. Today, this interest remains, but now the main role is given to analysts, who, armed with science and software, are trying to maintain a continuing interest at the proper level, giving sometimes quite good forecasts. Nowadays, the spectrum of approaches to the same problem has grown to unprecedented sizes: from trend lines and moving averages to neural networks and using several methods at once. It is not possible to describe all the approaches in one article. The article discusses basic preparation steps for forecasting, data correction, and common simple forecasting methods based on moving averages. Examples are given. R Studio, which is based on the R programming language and has a number of advantages and additional features, has been used for calculations and graphing, which greatly simplifies the work of the researcher. The article also provides code for graphing in a given programming language on an already configured environment and with attached libraries. In recent years, social media spread on the internet has burst into the lives of millions of people. That is why the visualization of relations of use of social media in Ukraine was made. Using the data on the population that uses the Internet in Ukraine, as well as on the data on the growth of Facebook users, a forecast was made with the growth of social network users in 5 years. A detailed description of the input and the results are provided. The issues that may be encountered and the impact of these problems on the final results and projections are also discussed.