What are the patterns of image diffusion during the Nepal Earthquake?
by Andrew Linenfelser, Caleb Hsu, Zhili Yang, Andrew Berumen

Questions

What is ratio of tweets in Nepali to tweets in Hindi? (Caleb Hsu)

Does user name: PrayForNepal have any original tweets or was the account created solely to repost tweets? (Andrew Linenfelser )

Who uploaded the most original tweets on 4/26, 4/27 and 4/28? (Zhili Yang)

Of the top images that got the most retweets what was the distribution of follower counts? (Andrey Shprengel)

Which users had the tweets closest to the Nepali Earthquake? (By Andrew Berumen)

# Discussion

According to our analysis, it appears that the majority of tweets collected in this dataset were tweeted in the morning (before 12:00PM Nepali Standard Time). This makes sense because although the earthquake itself occurred around noon, images likely did not begin circulating until the extend of the damage was evident and people had gone to safety. It would be interesting to see if image-less tweets were higher in volume near the actual time of the earthquake.

The three users that posted images closest to the actual time of the earthquake were agastya_kh, im_pritam3, and rmcroshan, and the users that created the most original image tweets were news agencies. Additionally, a number of the most retweeted images were originally created by users with a higher number of followers (with a few exceptions), so it is probable that these were the official twitters of various news agencies.

In terms of more specific difference in tweets, it appears the frequency of retweets in Hindi were 12% higher than retweets in Nepali, although this is also based off of a smaller sampling size of Hindi tweets (which likewise makes sense due to the demographics of Nepal). As with many natural disasters, Twitter accounts are created specifically for the purpose of retweeting news of the quake, as was the case with _prayfornepal_, which created no original retweets.