A DAILY MEASURE OF THE ITALIAN SENTIMENT ON THE ECONOMY BASED ON TWITTER DATA
Istat updates the Social Mood on Economy Index, the new experimental index first released in October 2018. The index provides daily measures of the Italian sentiment on the economy. These measures are derived from samples of public tweets in Italian captured in real time.
The production procedure of the index collects and processes only tweets containing at least one word belonging to a specific set of filter keywords, which has been designed by subject-matter experts. On average, this procedure elaborates through sentiment analysis techniques about 58,000 tweets per day.
The daily time series of the index is available in the Attachments section for the period 10th February 2016 – 31th December 2020.
The seasonal component of the series shows positive peaks in December and in the period late March early April, and a negative peak in the summer months.
After the robust growth at the end of September, in the fourth quarter of 2020 the Social Mood on Economy index showed a fluctuating trend. A phase of stability in the first half of October was followed by a decrease in the rest of the month. The first three weeks of November saw an upturn in the trend, followed by a more marked downturn that continued until Christmas. The month of December then closed with a stabilization of the trend.
This interactive plot shows the daily time series of the Social Mood on Economy Index (green line), along with the corresponding 15-days (blue line) and 30-days (red line) moving averages. The higher the value of the index, the better is the sentiment of the day. Significant peaks and valleys of the daily index have been annotated and are highlighted with a small square: just hover the mouse over the square and a tooltip will display the dominating topic(s) emerging from the tweets of the day (when several topics are listed, they are sorted by decreasing frequency).
This interactive plot shows the daily trend component of the Social Mood on Economy Index. The adopted seasonal adjustment methodology is described in a dedicated paragraph of the methodological note available in the Attachments section.
This interactive plot shows the time evolution of the number of tweets that have been collected and analyzed to compute the daily index. Discontinuities in the time series correspond either to downtime periods of the data collection system, or to the deletion of anomalous data generated by off-topic viral tweets (for details, please refer to the daily time series available in the Attachments section).
FOR FURTHER INFORMATION
Presentation at the 13th National conference of statistics – Istat (4 July 2018)
For information
Diego Zardetto
ph. +39 06 4673.4189
zardetto@istat.it
