Twitter habits in Spanish regions

Updated 10-09-2011 I’m sorry, unfortunately I made ​​a mistake in setting the GMT time to Spain. Data are 4 hours early. This changes the analysis a lot and I  will highlight the erroneous data in this post. Twitter is becoming more and more a part of people daily routine, from computers, tablets or smartphones, users follow their TL and publish tweets. Thanks to mobility, location is not a handicap, only to rest or to do priority tasks are an obstacle to using it. Therefore, analyzing the activity of their users we can understand their habits, allowing us to answer questions such as:

• What publication patterns are there in Spanish regions?
• When do people publish more in the morning or in the evening?
• What regions are early risers and which go to bed late?
• Is there a difference between the North and the South or between the mainland and the Islands?
• What is the pattern of the regions with large cities?

All these questions are answered in the following chart that shows the percentage of tweets posted every hour during the month of July in different regions of Spain. In this static image Madrid is highlighted but clicking on it you can access an interactive graphic and select other regions or a combination thereof. The graph shows a differentiated pattern of publication in the working week and at the weekend.

This image is 4 hours early

Conclusions:

• The sun never set on Twitter. Whenever users are tweeting all the time, night does not exist in Twitter world.
• Twitter has the same traffic pattern as phone calls. Seeing the tweet traffic curve it reminds me of the days when I worked with switching circuit because it resembles telephone call traffic. Twitter is just another channel to communicate with others. Although both, the working week and the weekend, peak publication is at 9:00 and at 19:00, the curves are somewhat different. During the working week it looks like a camel with two humps but at the weekend a third hump appears at 11:00. This lump may be due to the sleepers entering on the TL and if this hypothesis were true, people from Madrid are the sleepyheads in Spain.
• Morning vs. Evening: In the working week Madrid is different because there are more tweets in the morning than in the evening, but at the weekend it follows the same pattern as the other regions. Could it mean that in the working week from 18:00 to 20:00 all locals are in traffic jam? At lunchtime the Basque Country and the Northwest are those with the lowest frequency of publication, it could be logical because eating is not a trivial task in the north and it requires full concentration. In the evening the Northwest and the Canary Islands are those with most active tweeters.
• Early birds vs. Night owls.The earliest rise is Catalonia but is closely followed by Madrid, the Basque Country and the Canary Islands. The biggest revelers are Aragon and the Balearic Islands.
• North vs. South: In digital culture nothing is written. The Basque Country pattern is more similar to the South than other regions of the North.
• Islands vs. Peninsula: the islands can not be more different, the Canary Islands are early risers and the Balearic Islands are night owl. The Canary Islands are similar to the South and the Balearic Islands do not seem similar to any other region.
• Large cities: Regions with large cities are conditioned to the pattern of metropolis due to the volume of tweets generated.. Both, Madrid and Catalonia showed a similar profile, both are early risers and have less activity in the evening than other regions

Methodology:

• Data extracted from Twitter API since 1 to 31 of July of 2011, getting 32.878.322 tweets from 1.853.404 unique users, with this region distribution:
  • Aragon: 1.003.958 tweets from 46.219 unique users
  • Catalonia: 5.497.516 tweets from 284.585 unique users
  • Madrid: 9.045.342 tweets from 586.313 unique users
  • The Balearic Islands: 678.043 tweets from 41.318 unique users
  • The Basque Country: 1.138.830 tweets from 59.578 unique users
  • The Canary Islands: 871.765 tweets from 37.105 unique users
  • The East: 3.016.412 tweets from139. unique users
  • The Northwest: 5.137.426 tweets from 234.447 unique users
  • The South: 6.489.029 tweets from 231.576 unique users
• Tools:
  • Own datamining tools
  • Visualization made with Tableau Public