Research: Twitter usage pattern in Egypt (Statistics, Peaktimes and Analyses)

The research below answer the following questions

What is the usage pattern of Twitter by Egyptian users
What is the peak/low usage times

The Research

I managed to conduct this research on a 2 weeks span using a script which does a Twitter search and calculates the average time between tweets (ATBT) every 10 minutes, these ATBT numbers and other information were saved and analyzed to deduce the conclusion below

Note: the same research has been made on Facebook, check the link below
http://karim-ouda.blogspot.com/2011/11/research-facebook-usage-in-egypt.html

Context:

Tweets gathered using #Egypt search string
Research duration: from 24/10/2011 - 06/11/2011

Conclusion

- Top peak time (all days) is 14:40 PM

Note: i don't know why exactly the peak is @ 14:40 PM, if anyone have justification please let me know

- Peak hour (all days) is from 22:00-23:00PM

- Peak time spans are 14:00-16:00 PM and 20:00-23:00 PM

- Top day ( in terms of activity ) is Tuesday (Then Wednesday)

- The lowest (high ATBT) time span is from 1 AM and 10 AM

- The lowest (high ATBT) time is @ 5 AM

Charts

Overall Twitter usage by Egyptian users - Week #1 - ( 10 Min bases )

Overall Twitter usage by Egyptian users - Week #2 - ( 10 Min bases )

Overall Twitter usage by Egyptian users - Weekends

Overall Twitter usage by Egyptian users - Hourly

Technical Stuff

I am posting the code i created to conduct this research in-case you need to do the same on a different sample/purpose ( also check the license below )

Prerequisites

1- Linux Os
2- Basic linux/shell scripting/programming knowledge
3- 24/7 internet connection

Steps

1- Add the following Linux shell commands to a script file (file.sh)

### Do twitter search for "#Egypt", extract the timestamp of each tweet and save it to a temp file
rm -rf twitterTemp.txt; GET http://search.twitter.com/search.json?q=%23Egypt | grep -o -P "(?<=\"created_at\"\:)\".*?\"" | sed -e 's/$+0000\|"$//g' > twitterTemp.txt

### Read tmp file (containing the timestamps), calculate the average of the time between consecutive tweets, append info to a csv file
prevTimestamp=0;count=0; timestampsCount=`cat twitterTemp.txt | wc -l `; while read dateline ; do timestamp=`date -d "$dateline" +'%s'`; if [ $prevTimestamp -eq 0 ]; then prevTimestamp=$timestamp; fi ; count=$[$count+$[$prevTimestamp-$timestamp]]; prevTimestamp=$timestamp; done < twitterTemp.txt; avgTweetInterval=` echo "scale=1; $count/$timestampsCount" | bc ` ; echo `date "+%Y-%m-%d, %a, %H, %M"`, $avgTweetInterval >> twitter-statistics.csv ;

2- create a cron job to run this file every 10 min

crontab -e
0,10 * * * * /home/user/file.sh
CTRL-O
CTRL-X

FAQ

1- How do you track twitter usage pattern

Answer: through calculating the average of all time intervals between consecutive tweets

2- How do you know that the tweets were entered by Egyptian users

Answer: by searching the "#Egypt" hash tag, which is mostly used by Egyptian users

3- How do you calculate the average time between tweets (ATBT)

Answer: check the diagram below